Datasets:

tensorfeed
/

ai-ecosystem-daily

+ {"site":"tensorfeed.ai","url":"https://tensorfeed.ai","vertical":"AI infrastructure & news","description":"Real-time AI news aggregator, model + pricing tracker, infra status monitor, and agent registry.","adopted_at":"2026-04-30","manifest":"https://tensorfeed.ai/.well-known/agent-fair-trade.json","manifesto":"https://tensorfeed.ai/agent-fair-trade","receipt_key":"https://tensorfeed.ai/.well-known/tensorfeed-receipt-key.json","role":"standalone","notes":"Reference implementation of the standard."}

2026-10-12/agent-apis.jsonl ADDED Viewed

	@@ -0,0 +1,31 @@

+{"id":"tavily","name":"Tavily","vendor":"Tavily (joining Nebius)","category":"search","pricing":"1,000 free credits/mo; pay-as-you-go $0.008/credit; monthly Project plans from 4,000 credits","freeTier":"1,000 API credits/month, no credit card","hasMCP":true,"capabilities":["web search","AI-optimized snippets","extraction","research endpoint"],"url":"https://tavily.com","notes":"Default agent search API. AI-optimized result snippets and full-page extraction in one call. Tavily has announced it is joining Nebius; the API and pricing are unchanged so far."}
+{"id":"brave-search-api","name":"Brave Search API","vendor":"Brave","category":"search","pricing":"$5 per 1,000 requests with $5 in free credits every month","freeTier":"$5 in monthly credits (about 1,000 requests)","hasMCP":true,"capabilities":["web search","images","news","LLM context","custom reranking (Goggles)"],"url":"https://brave.com/search/api/","notes":"Independent web index (not Bing or Google). The old fixed free query tier is gone; every account now gets $5 in monthly credits applied automatically. Official MCP server available."}
+{"id":"exa","name":"Exa","vendor":"Exa Labs","category":"search","pricing":"Search $7 per 1,000 requests; contents $1 per 1,000 pages; Deep Search $12 to $15 per 1,000 requests","freeTier":"$20 credits on sign-up plus $10 in credits every month","hasMCP":true,"capabilities":["neural search","similarity","content extraction","deep research runs","people and company indexes"],"url":"https://exa.ai","notes":"Embedding-based semantic search. Better for \"find content like X\" queries than keyword search. Adds async Agent runs for deep research with structured, cited outputs."}
+{"id":"parallel","name":"Parallel Search API","vendor":"Parallel Web Systems","category":"search","pricing":"Pay as you go; Search API $0.001 to $0.005 per request for 10 results; Task API $0.005 to $2.40 per request","freeTier":"Up to 5,000 requests/month free, plus $5 in free credits monthly","hasMCP":true,"capabilities":["web search for agents","ranked URLs with compressed excerpts","extraction","deep research tasks","monitoring"],"url":"https://parallel.ai","notes":"Web search and research APIs built for AI agents rather than humans: search returns ranked URLs with token-efficient excerpts, and the Task API runs deep research with structured outputs. SOC 2 compliant, with an official MCP server and CLI."}
+{"id":"serpapi","name":"SerpAPI","vendor":"SerpAPI","category":"search","pricing":"Developer $75/mo for 5,000 searches; Production $150/mo for 15,000","freeTier":"250 searches/month","hasMCP":true,"capabilities":["Google SERP","Bing","YouTube","images"],"url":"https://serpapi.com","notes":"Scrapes real Google, Bing, and other result pages. Use when you need actual SERP data rather than a curated index. Offers a U.S. Legal Shield and an MCP integration."}
+{"id":"firecrawl","name":"Firecrawl","vendor":"Firecrawl","category":"web-scraping","pricing":"1,000 credits/mo free; Hobby $19/mo ($16 billed annually) for 5,000 credits","freeTier":"1,000 credits/month (500 searches or 1,000 pages)","hasMCP":true,"capabilities":["scrape","crawl","search","JS rendering","structured extraction"],"url":"https://firecrawl.dev","notes":"Designed for AI agents. Returns clean markdown and supports schema-based extraction. Credits are shared across endpoints, so searches and scraped pages draw on one balance."}
+{"id":"jina-reader","name":"Jina Reader","vendor":"Jina AI (Elastic)","category":"web-scraping","pricing":"Free without a key; token-based pricing with an API key","freeTier":"20 RPM without a key; 500 RPM and 10M free tokens with a new key","hasMCP":true,"capabilities":["URL to markdown","web search (s.jina.ai)","JS rendering","screenshot mode"],"url":"https://jina.ai/reader/","notes":"Prepend r.jina.ai/ to any URL for agent-ready content; s.jina.ai searches the web and returns LLM-friendly results. Official MCP server at mcp.jina.ai. Elastic has sold the Jina On-Prem commercial license for the reader models as its own SKU since August 10, 2026."}
+{"id":"browserbase","name":"Browserbase","vendor":"Browserbase","category":"web-scraping","pricing":"Free, $20/mo, and $99/mo plans; custom above","freeTier":"3 concurrent browsers, 1 browser hour, 1,000 search and 1,000 fetch calls","hasMCP":true,"capabilities":["headless browsers for agents","search and fetch APIs","managed browser agents","identity and observability"],"url":"https://www.browserbase.com","notes":"Cloud browser infrastructure for agents that need a real browser instead of a scraper. Also ships web search and fetch APIs, managed agent runs, and the open-source Stagehand framework, with an official MCP server."}
+{"id":"apify","name":"Apify","vendor":"Apify","category":"web-scraping","pricing":"Free plan with $5/mo platform credit; Starter $19/mo plus usage; $0.20 per compute unit","freeTier":"$5 credit/month, no credit card","hasMCP":true,"capabilities":["headless browser","pre-built actors","scheduled crawls","x402 pay-per-run"],"url":"https://apify.com","notes":"Marketplace of pre-built scrapers (\"Actors\"). Best for scraping specific named sites without writing code. Agents can now run Actors and pay per run over x402."}
+{"id":"scrapingbee","name":"ScrapingBee","vendor":"ScrapingBee","category":"web-scraping","pricing":"Hobby $19/mo for 75,000 credits; Freelance $49/mo for 250,000","freeTier":"1,000 free credits, no credit card","hasMCP":true,"capabilities":["proxy rotation","JS rendering","AI data extraction","dedicated search and site APIs"],"url":"https://www.scrapingbee.com","notes":"Proxy-rotating headless scraper. Good when target sites have anti-bot defenses. Now offers an MCP server and dedicated APIs for Google, Amazon, YouTube, and Walmart."}
+{"id":"openweather","name":"OpenWeather","vendor":"OpenWeather","category":"weather","pricing":"Free plan; One Call API 4.0 is pay per call above 1,000/day; fixed monthly plans from Startup to Expert","freeTier":"60 calls/min and 1M calls/month on the free plan; One Call 4.0 includes 1,000 calls/day free","hasMCP":false,"capabilities":["current","forecast","historical","air quality","alerts"],"url":"https://openweathermap.org/api","notes":"Most-used weather API. One Call API 4.0 (timeline-based) is now the flagship endpoint, with free permanent access for low-volume use."}
+{"id":"noaa-nws","name":"NWS API","vendor":"NOAA / NWS","category":"weather","pricing":"Free","freeTier":"Unlimited (with rate limits)","hasMCP":false,"capabilities":["US-only forecasts","alerts","observations"],"url":"https://www.weather.gov/documentation/services-web-api","notes":"Free US government weather API. No API key required. US coverage only; for global use OpenWeather."}
+{"id":"tomorrow-io","name":"Tomorrow.io","vendor":"Tomorrow.io","category":"weather","pricing":"Free tier; usage-based after","freeTier":"500 requests/day","hasMCP":false,"capabilities":["hyperlocal forecast","hourly","60+ data fields"],"url":"https://www.tomorrow.io/weather-api/","notes":"Highest-resolution weather API. Useful for agents doing logistics, agriculture, or event planning."}
+{"id":"alpha-vantage","name":"Alpha Vantage","vendor":"Alpha Vantage","category":"finance","pricing":"Free tier (25 requests/day); premium from $49.99/mo for 75 requests/min","freeTier":"25 requests/day","hasMCP":true,"capabilities":["stocks","forex","crypto","technical indicators","fundamental data","news sentiment"],"url":"https://www.alphavantage.co","notes":"Most-used free stock API. Generous free tier; rate limits tighten at scale. Ships an official MCP server for chat and coding agents."}
+{"id":"polygon","name":"Massive (formerly Polygon.io)","vendor":"Massive","category":"finance","pricing":"Stocks Basic free; Stocks Starter $29/mo; partner datasets $99/mo each","freeTier":"5 calls/min, end-of-day data, 2 years history","hasMCP":true,"capabilities":["real-time stocks","options","forex","crypto","futures","WebSocket streams"],"url":"https://massive.com","notes":"Production-grade market data. Polygon.io rebranded to Massive on October 27, 2025 with existing APIs and keys unchanged. WebSocket support is the main differentiator from Alpha Vantage, and a remote MCP server is now available."}
+{"id":"coingecko","name":"CoinGecko","vendor":"CoinGecko","category":"finance","pricing":"Free Demo plan; Basic $35/mo and Analyst $129/mo (billed yearly)","freeTier":"Demo plan: 10,000 call credits/month, attribution required","hasMCP":true,"capabilities":["crypto prices","market cap","historical","NFTs","on-chain DEX data"],"url":"https://www.coingecko.com/en/api","notes":"The default crypto data API, tracking more than 21,000 coins. Free Demo tier is enough for most agent workflows, plus a keyless public API and an official MCP server."}
+{"id":"google-maps","name":"Google Maps Platform","vendor":"Google","category":"maps","pricing":"Pay per request, with free monthly usage caps per SKU","freeTier":"Free usage cap on every SKU each month (replaced the $200 monthly credit on March 1, 2025)","hasMCP":true,"capabilities":["geocoding","directions","places","distance matrix","street view","grounding for LLMs"],"url":"https://mapsplatform.google.com","notes":"Most-used maps API. Free usage caps cover most small-agent use cases. Grounding Lite brings Maps data to any LLM over MCP."}
+{"id":"mapbox","name":"Mapbox","vendor":"Mapbox","category":"maps","pricing":"Free tier; pay-as-you-go","freeTier":"Free monthly allowances on most APIs (for example 100,000 Directions requests)","hasMCP":true,"capabilities":["geocoding","directions","static maps","tiles","isochrones"],"url":"https://www.mapbox.com","notes":"Generous free tier and better visual customization than Google. Official Mapbox MCP server plus agent skills for location AI."}
+{"id":"osrm","name":"OSRM","vendor":"OSRM (open-source)","category":"maps","pricing":"Free (self-host)","freeTier":"Unlimited self-host","hasMCP":false,"capabilities":["routing","distance matrix","isochrones"],"url":"http://project-osrm.org","notes":"Open-source routing engine. Self-host for unlimited usage. Public demo server for testing only."}
+{"id":"resend","name":"Resend","vendor":"Resend","category":"email","pricing":"Free tier; Pro $20/mo for 50,000 emails; Scale $90/mo for 100,000","freeTier":"3,000 emails/month, 100/day","hasMCP":true,"capabilities":["transactional email","React email templates","webhooks","marketing emails"],"url":"https://resend.com","notes":"Developer-first email API with React-based templates and an official MCP server. The default for new SaaS and agent stacks."}
+{"id":"postmark","name":"Postmark","vendor":"ActiveCampaign","category":"email","pricing":"Basic $15/mo for 10,000 emails; Pro from $16.50/mo","freeTier":"100 emails/month","hasMCP":true,"capabilities":["transactional email","high deliverability","streams","webhooks"],"url":"https://postmarkapp.com","notes":"Best-in-class deliverability for transactional email. ActiveCampaign maintains an official Postmark MCP server."}
+{"id":"twilio-sms","name":"Twilio SMS","vendor":"Twilio","category":"sms","pricing":"Pay-per-message; US SMS from $0.0083","freeTier":"Free trial for new accounts","hasMCP":true,"capabilities":["SMS","MMS","RCS","WhatsApp","voice"],"url":"https://www.twilio.com/messaging","notes":"The most-used programmable messaging API. WhatsApp and RCS included. MCP server available from Twilio Labs. Carrier fees apply on top of per-message rates."}
+{"id":"stripe","name":"Stripe","vendor":"Stripe","category":"payments","pricing":"2.9% + 30¢ per successful domestic card transaction","freeTier":"No setup or monthly fees","hasMCP":true,"capabilities":["payments","subscriptions","invoicing","connect (multi-party)"],"url":"https://stripe.com","notes":"Default agent-to-human payments API. Official Stripe MCP server and agent toolkit with restricted-key auth."}
+{"id":"lemon-squeezy","name":"Lemon Squeezy","vendor":"Stripe","category":"payments","pricing":"5% + 50¢ per transaction","freeTier":"No monthly fees","hasMCP":false,"capabilities":["merchant of record","EU VAT","subscriptions","license keys"],"url":"https://www.lemonsqueezy.com","notes":"Merchant of record that handles VAT and sales tax for you. Owned by Stripe; a 2026 update pairs it with Stripe Managed Payments. Good fit for solo developers selling internationally."}
+{"id":"e2b","name":"E2B","vendor":"E2B","category":"code-execution","pricing":"Hobby free plus usage; Pro $150/mo plus usage billed by the second","freeTier":"$100 one-time usage credit, 20 concurrent sandboxes, 1-hour sessions","hasMCP":false,"capabilities":["secure sandboxes","Jupyter","long-running","Linux"],"url":"https://e2b.dev","notes":"Cloud sandboxes designed for AI code execution and an OpenAI Agents API sandbox partner. Its standalone MCP server repo is now deprecated."}
+{"id":"daytona","name":"Daytona","vendor":"Daytona","category":"code-execution","pricing":"Pay-as-you-go; vCPU $0.0504/hr, memory $0.0162/GiB-hr, GPUs from $0.57/hr","freeTier":"$200 in free compute","hasMCP":true,"capabilities":["sandboxes in milliseconds","Linux + Windows","GPU sandboxes","SDKs in TypeScript, Python, Ruby, Go, Java"],"url":"https://www.daytona.io","notes":"Fast-starting sandboxes for agent code execution, including GPU and Windows options. An OpenAI Agents API sandbox partner, with an official MCP server."}
+{"id":"modal","name":"Modal","vendor":"Modal Labs","category":"code-execution","pricing":"Starter $0 plus compute; billed per second","freeTier":"$30 monthly compute credit","hasMCP":false,"capabilities":["serverless GPU + CPU","secrets","volumes","web endpoints"],"url":"https://modal.com","notes":"Serverless code execution including GPU. Good for agents that need heavy compute on demand, and an OpenAI Agents API sandbox partner."}
+{"id":"replit-agent","name":"Replit Agent","vendor":"Replit","category":"code-execution","pricing":"Core $20/mo; Pro $100/mo with 10 parallel agents; Enterprise custom","freeTier":"Free Replit account","hasMCP":false,"capabilities":["IDE-integrated execution","Linux containers","multi-language"],"url":"https://replit.com","notes":"Ephemeral execution sandboxes inside Replit. Less precise than E2B for agent control, but bundled with the Replit subscription."}
+{"id":"unstructured","name":"Unstructured","vendor":"Unstructured.io","category":"ocr","pricing":"Open-source library free; hosted pay-as-you-go $0.015/page after 10,000 free pages","freeTier":"10,000 free pages on the hosted platform","hasMCP":false,"capabilities":["PDF parsing","OCR","tables","images","60+ formats"],"url":"https://unstructured.io","notes":"The default RAG ingest pipeline. Open-source library handles most formats; the hosted platform adds higher accuracy and scale."}
+{"id":"llamaparse","name":"LlamaParse","vendor":"LlamaIndex","category":"ocr","pricing":"Credit-based; Starter $50/mo with 40K credits, Pro $500/mo with 400K credits","freeTier":"10,000 credits/month","hasMCP":false,"capabilities":["PDF parsing","tables","figures","multimodal","extraction"],"url":"https://www.llamaindex.ai/llamaparse","notes":"GenAI-native document parser at the core of the LlamaIndex document processing platform, billed in credits across parsing and extraction."}
+{"id":"reducto","name":"Reducto","vendor":"Reducto","category":"ocr","pricing":"Standard pay-as-you-go after free usage; Growth custom","freeTier":"$150 in free usage","hasMCP":false,"capabilities":["PDF parsing","forms","tables","extract, edit, and split APIs"],"url":"https://reducto.ai","notes":"Premium document parser. Highest accuracy on complex tables; targets enterprise RAG and compliance use cases. Offers up to $5,000 in migration credits."}

2026-10-12/agent-provisioning.jsonl ADDED Viewed

	@@ -0,0 +1,22 @@

+{"id":"cloudflare","name":"Cloudflare","vendor":"Cloudflare","category":"cdn-edge","status":"live","role":"Edge hosting, Workers, databases, storage, queues, browser rendering, domain registration, AI","defaultMonthlyCap":100,"notes":"Co-author of the protocol with Stripe. Account creation, domain registration through Cloudflare Registrar, billing, and Workers deploy are all agent-callable, and Cloudflare publishes its service catalog as a JSON REST API. Any platform with signed-in users can integrate with Cloudflare the same way Stripe does. The reference implementation.","url":"https://www.cloudflare.com"}
+{"id":"vercel","name":"Vercel","vendor":"Vercel","category":"hosting","status":"live","role":"Frontend hosting, serverless functions, Next.js platform","defaultMonthlyCap":100,"notes":"In the catalog as vercel/project. Project creation and paid plan upgrades run through Stripe Projects. Strongest Next.js and AI SDK story for agent-built apps, and also an OpenAI Agents API sandbox partner.","url":"https://vercel.com"}
+{"id":"netlify","name":"Netlify","vendor":"Netlify","category":"hosting","status":"live","role":"Frontend hosting, serverless functions, edge functions","defaultMonthlyCap":100,"notes":"Not on our May 3 list but now in the catalog as netlify/project, so agents picking a frontend host can choose it alongside Vercel.","url":"https://www.netlify.com"}
+{"id":"fly","name":"Fly.io","vendor":"Fly.io","category":"hosting","status":"live","role":"Global app hosting, full-VM workloads, multi-region","defaultMonthlyCap":100,"notes":"Not on our May 3 list but now in the catalog as flyio/app, covering hosting and databases. The strongest global-region option for non-Next.js workloads that agents can provision.","url":"https://fly.io"}
+{"id":"railway","name":"Railway","vendor":"Railway","category":"hosting","status":"live","role":"App hosting, databases, cache, storage","defaultMonthlyCap":100,"notes":"In the catalog as railway/hosting, spanning hosting, databases, cache, and storage. Usage-based billing by the second with a free tier, which suits agents that spin services up and down.","url":"https://railway.com"}
+{"id":"render","name":"Render","vendor":"Render","category":"hosting","status":"live","role":"Static sites and app hosting, managed databases","defaultMonthlyCap":100,"notes":"In the catalog as render/static-site, with hosting and database categories. Another full-stack host agents can self-serve onto without a Vercel or Netlify account.","url":"https://render.com"}
+{"id":"supabase","name":"Supabase","vendor":"Supabase","category":"database","status":"live","role":"Postgres database, auth, storage, realtime","defaultMonthlyCap":100,"notes":"In the catalog as supabase/project, spanning database, auth, and storage. Pairs well with Vercel for Next.js stacks.","url":"https://supabase.com"}
+{"id":"planetscale","name":"PlanetScale","vendor":"PlanetScale","category":"database","status":"live","role":"Managed Postgres and MySQL, branching, schema migrations","defaultMonthlyCap":100,"notes":"In the catalog as planetscale/postgresql, so the agent-provisionable service is PlanetScale Postgres. Branch-based schema workflows stay agent-driveable.","url":"https://planetscale.com"}
+{"id":"neon","name":"Neon","vendor":"Neon","category":"database","status":"live","role":"Serverless Postgres, branching, instant cold starts, auth","defaultMonthlyCap":100,"notes":"Not on our May 3 list but now in the catalog as neon/postgres, covering database and auth. Agents wiring a database from scratch now have three Postgres options: Supabase, PlanetScale, and Neon.","url":"https://neon.tech"}
+{"id":"turso","name":"Turso","vendor":"Turso","category":"database","status":"live","role":"Edge SQLite (libSQL), per-user databases","defaultMonthlyCap":100,"notes":"Not on our May 3 list but now in the catalog as turso/database. Fills the edge-first SQLite gap alongside the Cloudflare, Vercel, and Supabase core.","url":"https://turso.tech"}
+{"id":"upstash","name":"Upstash","vendor":"Upstash","category":"database","status":"live","role":"Serverless Redis, messaging, vector store, search","defaultMonthlyCap":100,"notes":"Unconfirmed in our May 3 reading; now confirmed in the catalog as upstash/redis, spanning cache, database, AI, messaging, and search.","url":"https://upstash.com"}
+{"id":"clerk","name":"Clerk","vendor":"Clerk","category":"auth","status":"live","role":"Auth, sessions, user management, organizations","defaultMonthlyCap":100,"notes":"In the catalog as clerk/auth. Application creation and OAuth provider configuration are agent-callable. No longer the only auth option: Auth0, WorkOS, Neon, and Privy have joined.","url":"https://clerk.com"}
+{"id":"auth0","name":"Auth0","vendor":"Okta","category":"auth","status":"live","role":"Auth, SSO, identity-as-a-service","defaultMonthlyCap":100,"notes":"The largest auth provider in the market was not on our May 3 list and is now in the catalog as Auth0/Okta (auth0/client), so agent-driven sign-ups no longer default to Clerk.","url":"https://auth0.com"}
+{"id":"workos","name":"WorkOS","vendor":"WorkOS","category":"auth","status":"live","role":"Enterprise SSO, SCIM, directory sync, RBAC, audit logs","defaultMonthlyCap":100,"notes":"Now in the catalog as workos/auth, pitched as infrastructure for selling to enterprise: SSO, SCIM, RBAC, audit logs, and AI governance.","url":"https://workos.com"}
+{"id":"sentry","name":"Sentry","vendor":"Sentry","category":"observability","status":"live","role":"Error tracking, performance monitoring","defaultMonthlyCap":100,"notes":"In the catalog as sentry/project. Project creation and SDK key issuance are API-driveable. Datadog and Pydantic Logfire are also in the observability lineup.","url":"https://sentry.io"}
+{"id":"posthog","name":"PostHog","vendor":"PostHog","category":"observability","status":"live","role":"Product analytics, feature flags, session replay","defaultMonthlyCap":100,"notes":"In the catalog as posthog/analytics, spanning analytics, AI, feature flags, and observability. Often paired with Sentry for full observability.","url":"https://posthog.com"}
+{"id":"inngest","name":"Inngest","vendor":"Inngest","category":"background-jobs","status":"live","role":"Durable workflows, background jobs, event-driven functions","defaultMonthlyCap":100,"notes":"In the catalog as inngest/app under the queue category. App creation and event and function registration are agent-driveable.","url":"https://www.inngest.com"}
+{"id":"openrouter","name":"OpenRouter","vendor":"OpenRouter (joining Stripe)","category":"ai-infrastructure","status":"live","role":"Unified LLM API and inference routing across model providers","defaultMonthlyCap":100,"notes":"In the catalog as openrouter/api, giving agents one API across many model providers. Stripe agreed to acquire OpenRouter on August 19, 2026, putting the billing rail and the inference gateway under one owner.","url":"https://openrouter.ai"}
+{"id":"agentmail","name":"AgentMail","vendor":"AgentMail","category":"email","status":"live","role":"Email inboxes for AI agents","defaultMonthlyCap":100,"notes":"In the catalog as agentmail/api. Gives each agent its own email inbox through an API, the email option built for agents rather than for human senders. Customer.io and Twilio also list email services.","url":"https://www.agentmail.to"}
+{"id":"huggingface","name":"Hugging Face","vendor":"Hugging Face (NVIDIA acquisition pending)","category":"ai-infrastructure","status":"pending","role":"Model hosting, inference endpoints, Spaces, dataset hosting","defaultMonthlyCap":100,"notes":"Tracked as a launch partner in our May 3 reading, but it does not appear in the public 63-provider Stripe Projects catalog as of 2026-09-14. NVIDIA agreed to acquire Hugging Face for about $12.9 billion on September 3, 2026. Until it lists, OpenRouter is the AI inference service agents can provision through Stripe Projects.","url":"https://huggingface.co"}
+{"id":"stytch","name":"Stytch","vendor":"Stytch","category":"auth","status":"pending","role":"Auth, passwordless, biometrics, M2M","defaultMonthlyCap":100,"notes":"Developer-first auth with strong M2M primitives that should map cleanly to agents, but still not in the catalog while Auth0, WorkOS, and Clerk are.","url":"https://stytch.com"}
+{"id":"resend","name":"Resend","vendor":"Resend","category":"email","status":"pending","role":"Transactional email, React-based templates","defaultMonthlyCap":100,"notes":"Unconfirmed in our May 3 reading; the public catalog now shows it is not listed. The default email API for many agent stacks, so agents provisioning email through Stripe Projects route to AgentMail, Customer.io, or Twilio instead.","url":"https://resend.com"}

2026-10-12/agents-activity.jsonl ADDED Viewed

	@@ -0,0 +1,50 @@

+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:15:34.088Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:15:34.089Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:16:39.700Z"}
+{"bot":"Unknown Bot","endpoint":"/feed.json","timestamp":"2026-10-12T17:16:08.857Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:15:40.058Z"}
+{"bot":"python-requests","endpoint":"/api/premium/research/citation-velocity","timestamp":"2026-10-12T17:15:33.384Z"}
+{"bot":"python-requests","endpoint":"/api/premium/policy/timeline","timestamp":"2026-10-12T17:14:49.808Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:14:04.011Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:14:04.011Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:14:39.695Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:12:33.957Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:13:39.695Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:12:39.729Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:11:03.901Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:11:03.902Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:11:39.693Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:12:33.955Z"}
+{"bot":"python-requests","endpoint":"/api/premium/research/emerging-keywords","timestamp":"2026-10-12T17:12:02.669Z"}
+{"bot":"Unknown Crawler","endpoint":"/api/premium/routing","timestamp":"2026-10-12T17:11:41.111Z"}
+{"bot":"python-requests","endpoint":"/api/mcp","timestamp":"2026-10-12T17:10:44.301Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:09:33.848Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:10:39.700Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:09:39.857Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:08:03.759Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:08:03.761Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:08:39.695Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:09:33.846Z"}
+{"bot":"python-requests","endpoint":"/api/premium/inference-providers/arbitrage","timestamp":"2026-10-12T17:09:14.969Z"}
+{"bot":"python-requests","endpoint":"/api/mcp","timestamp":"2026-10-12T17:08:13.980Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:06:33.708Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:07:39.696Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:06:39.994Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:05:03.648Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:05:03.649Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:05:39.694Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:06:33.707Z"}
+{"bot":"Unknown Spider","endpoint":"/_next/static/chunks/webpack-0b2f7d6b1b54e886.js","timestamp":"2026-10-12T17:06:30.030Z"}
+{"bot":"Unknown Spider","endpoint":"/_next/static/css/0e213492a891d9d9.css","timestamp":"2026-10-12T17:06:21.815Z"}
+{"bot":"Unknown Spider","endpoint":"/about","timestamp":"2026-10-12T17:06:20.437Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:03:33.595Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:04:39.700Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:03:40.031Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:02:03.534Z"}
+{"bot":"Unknown Spider","endpoint":"/api/breaking","timestamp":"2026-10-12T17:02:03.536Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status/summary","timestamp":"2026-10-12T17:02:39.694Z"}
+{"bot":"Unknown Spider","endpoint":"/api/status","timestamp":"2026-10-12T17:03:33.594Z"}
+{"bot":"python-requests","endpoint":"/api/premium/whats-new","timestamp":"2026-10-12T17:03:05.301Z"}
+{"bot":"Unknown Agent","endpoint":"/api/mcp","timestamp":"2026-10-12T17:01:59.367Z"}
+{"bot":"Unknown Agent","endpoint":"/api/mcp","timestamp":"2026-10-12T17:01:58.969Z"}
+{"bot":"Unknown Bot","endpoint":"/feed.xml","timestamp":"2026-10-12T17:01:31.551Z"}

2026-10-12/agents-directory.jsonl ADDED Viewed

	@@ -0,0 +1,28 @@

+{"id":"muse-code","name":"Muse Code","provider":"Meta","category":"coding","description":"A terminal coding agent released August 5, 2026 and co-trained as one unit with Muse Spark so the model and the harness share goals, compaction, and subagent recipes. It left beta on August 31, 2026 and moved to Muse Spark 1.3 on September 2, 2026. Meta documents it as running on macOS, Linux, and Windows from one codebase; it runs persistent background agents across a session and keeps a replay-safe event log so it survives a crash.","url":"https://dev.meta.ai","pricing":"Flat monthly plans at $5 (Everyday Usage), $15 (High Usage), and $50 (Power Usage), or usage-based via the Meta Model API: $1.25/$4.25 per 1M tokens standard, or $0.10/$0.20 on the contributor tier if you let Meta train on your data","launched":2026}
+{"id":"deepseek-harness","name":"DeepSeek Harness","provider":"DeepSeek","category":"coding","description":"A plugin-first open-source coding agent released August 13, 2026 under MIT, aimed directly at Claude Code. Shipped the same day DeepSeek moved V4-Pro to general availability and raised API prices, pairing an open harness with a more expensive model. Still labeled a developer preview with breaking changes expected, and past 238,000 GitHub stars.","url":"https://github.com/deepseek-ai/deepseek-harness","pricing":"Free and open source under MIT; model inference billed separately","launched":2026}
+{"id":"grok-build","name":"Grok Build","provider":"xAI (SpaceXAI)","category":"coding","description":"xAI's official terminal coding agent, invoked as the grok command. Delegates larger tasks to parallel subagents that can each run in their own git worktree, and has shipped Grok 4.7, with a 500K context window, since the model's release on September 21, 2026.","url":"https://x.ai/build","pricing":"Free to try within limits; higher usage comes with SuperGrok ($30/mo), SuperGrok Plus ($100/mo), and SuperGrok Heavy. API use bills at model rates, currently $2/$6 per 1M tokens for Grok 4.7 on prompts under 200K tokens, doubling to $4/$12 above that","launched":2026}
+{"id":"openai-codex","name":"Codex","provider":"OpenAI","category":"coding","description":"OpenAI's coding agent, available in the ChatGPT desktop app, on the web, in the open-source Codex CLI, in IDE extensions, and in Codex cloud, where it completes tasks end to end, from routine pull requests to hard engineering problems. Its harness is the one OpenAI now exposes to developers through the Agents API.","url":"https://openai.com/codex/","pricing":"Included with ChatGPT plans: Free, Go $8/mo, Plus $20/mo, Pro from $100/mo, plus Business and Enterprise seats","launched":2025}
+{"id":"google-antigravity","name":"Google Antigravity","provider":"Google","category":"coding","description":"Google's agent-first development platform, introduced November 18, 2025 alongside Gemini 3 and now generally available. Pairs a familiar AI IDE with a manager surface where agents plan, code, and verify work across the editor, terminal, and browser, with model choice spanning Gemini, Claude, and gpt-oss.","url":"https://antigravity.google","pricing":"Free for individuals with weekly rate limits; higher limits with Google AI Pro and Google AI Ultra","launched":2025}
+{"id":"claude-code","name":"Claude Code","provider":"Anthropic","category":"coding","description":"An agentic coding tool that lets Claude operate directly in your terminal, IDE, desktop app, or browser, reading files, editing code, running commands, and managing git workflows autonomously.","url":"https://code.claude.com/docs","pricing":"Included with Claude Pro ($20/mo), Max (from $100/mo), Team, and Enterprise plans; also usage-based via the Claude API","launched":2025}
+{"id":"cursor","name":"Cursor","provider":"SpaceXAI (formerly Anysphere)","category":"coding","description":"An AI-native code editor built on VS Code that provides inline code generation, multi-file editing, and codebase-aware chat powered by multiple foundation models. SpaceX closed its $60 billion all-stock acquisition of maker Anysphere on August 14, 2026, folding Cursor into the SpaceXAI unit alongside Grok Build and Grok Bot.","url":"https://cursor.com","pricing":"Hobby free, Pro $20/mo, Pro+ $60/mo, Ultra $200/mo, Teams $40/user/mo (Premium seats $120/user/mo), Enterprise custom","launched":2023}
+{"id":"github-copilot","name":"GitHub Copilot","provider":"GitHub / Microsoft","category":"coding","description":"An AI pair programmer integrated into popular editors that suggests code completions, generates functions from comments, and offers chat-based coding assistance.","url":"https://github.com/features/copilot","pricing":"Free tier, Pro $10/mo, Pro+ $39/mo, Max $100/mo, Business $19/seat/mo, Enterprise $39/seat/mo; usage metered in GitHub AI Credits","launched":2022}
+{"id":"windsurf","name":"Devin Desktop (formerly Windsurf)","provider":"Cognition","category":"coding","description":"An AI-native developer environment (rebranded from Windsurf in June 2026) that provides a unified surface for managing local and cloud coding agents, including Devin, Claude Code, Codex, and custom agents through the Agent Client Protocol.","url":"https://devin.ai","pricing":"Free tier, Pro $20/mo, Max $200/mo, Teams $80/mo plus $40/mo per full seat, Enterprise custom","launched":2024}
+{"id":"perplexity","name":"Perplexity","provider":"Perplexity AI","category":"research","description":"An AI-powered answer engine that searches the web in real time, synthesizes information from multiple sources, and provides cited responses to complex questions.","url":"https://www.perplexity.ai","pricing":"Free tier, Pro $17/mo and Max $167/mo when billed annually","launched":2022}
+{"id":"elicit","name":"Elicit","provider":"Elicit Inc.","category":"research","description":"A research assistant that helps find and analyze academic papers, extract key claims, and summarize findings across large bodies of scientific literature.","url":"https://elicit.com","pricing":"Free Basic tier; Pro $49/user/mo ($588 billed annually) and Scale $169/user/mo ($2,028 billed annually); Enterprise custom","launched":2021}
+{"id":"consensus","name":"Consensus","provider":"Consensus NLP","category":"research","description":"A search engine that uses AI to find and synthesize results from peer-reviewed scientific research, providing evidence-based answers with citations.","url":"https://consensus.app","pricing":"Free tier, Pro $20/mo ($144 billed annually); a higher Deep tier and custom Team plans are also offered","launched":2022}
+{"id":"chatgpt","name":"ChatGPT","provider":"OpenAI","category":"general","description":"A general-purpose AI assistant that can handle conversation, writing, coding, analysis, and web browsing, with agent mode, apps, custom GPTs, and the Codex coding agent built in on paid plans.","url":"https://chatgpt.com","pricing":"Free tier, Go $8/mo, Plus $20/mo, Pro $100 or $200/mo (new $200 sign-ups paused since September 10, 2026), Business $25/seat/mo ($20 billed annually), Enterprise custom","launched":2022}
+{"id":"gemini","name":"Gemini","provider":"Google","category":"general","description":"Google's multimodal AI assistant with deep integration into Google Workspace, Search, and Android, capable of handling text, images, code, and long documents.","url":"https://gemini.google.com","pricing":"Free tier, Google AI Plus $4.99/mo, Google AI Pro $19.99/mo, Google AI Ultra from $99.99/mo","launched":2023}
+{"id":"gemini-spark","name":"Gemini Spark","provider":"Google","category":"general","description":"A 24/7 personal AI agent announced at Google I/O May 2026 that works in the background even when your phone and laptop are off, taking actions across Gmail, Calendar, Drive, Docs, and other connected apps under the user's direction. Now rolled out to paying Google AI subscribers in select countries and to select business users.","url":"https://gemini.google/overview/agent/spark/","pricing":"Included with Google AI Pro ($19.99/mo) and Google AI Ultra (from $99.99/mo) for users 18 and over in select countries","launched":2026}
+{"id":"openclaw","name":"OpenClaw","provider":"OpenClaw Foundation","category":"general","description":"Open-source personal AI agent that runs locally, keeps context across conversations, and acts on the user's own machine, driven mainly through messaging platforms rather than a chat window. Shipped in 2025 as Clawdbot, renamed Moltbot over a trademark conflict, and settled on OpenClaw in January 2026; now stewarded by the independent OpenClaw Foundation, a 501(c)(3) with no paid tier, hosted service, or token.","url":"https://github.com/openclaw/openclaw","pricing":"Free and open source; you pay only whatever model provider it calls","launched":2025}
+{"id":"claude-cowork","name":"Claude Cowork","provider":"Anthropic","category":"general","description":"Anthropic's agent for knowledge work beyond coding, launched January 12, 2026 as a way to bring Claude Code's agentic loop to the rest of your work. It reads local files, connected tools, and the web, then produces finished documents, spreadsheets, and decks; scheduled tasks keep running unattended after the laptop closes. Generally available in the Claude desktop app on macOS and Windows, with web and mobile access added since.","url":"https://claude.com/product/cowork","pricing":"Included with Claude Pro ($20/mo), Max (from $100/mo), Team, and Enterprise plans","launched":2026}
+{"id":"grok-bot","name":"Grok Bot","provider":"xAI (SpaceXAI)","category":"general","description":"Always-on agents that each get their own cloud computer, sign into a customer's existing tools, and finish multi-step jobs unsupervised, including inside apps and sites that expose no clean API or MCP integration. Bots run in parallel around the clock after the user closes the laptop, run routines on a schedule, and come back when they need approval rather than direction at every step. A macOS desktop app is available.","url":"https://x.ai/bot","pricing":"No standalone plan. Included with weekly usage on SuperGrok ($30/mo) and higher xAI plans, Cursor Pro ($20/mo), Pro+, and Ultra, and Cursor Teams ($40/seat/mo)","launched":2026}
+{"id":"manus","name":"Manus","provider":"Butterfly Effect","category":"general","description":"An autonomous general agent that plans and runs multi-step work on its own cloud computer, covering research, data analysis, slides, sites, and spreadsheets, and launched in invite-only beta on March 6, 2025. Meta bought maker Butterfly Effect for a reported $2 billion in December 2025, China's NDRC ordered the deal unwound in April 2026, and Manus resumed independent operations on September 1, 2026.","url":"https://manus.im","pricing":"Free tier with 300 daily refresh credits; Pro $20/mo (from 4,000 monthly credits) or $40/mo (from 8,000); Team $20/seat/mo; 17 percent off with annual billing","launched":2025}
+{"id":"claude","name":"Claude","provider":"Anthropic","category":"general","description":"A helpful AI assistant known for nuanced instruction-following, long-context understanding, and careful reasoning across writing, analysis, coding, and research tasks.","url":"https://claude.ai","pricing":"Free tier, Pro $20/mo ($17/mo billed annually), Max from $100/mo, Team $25/seat/mo ($20 billed annually), Enterprise $20/seat plus usage at API rates","launched":2023}
+{"id":"midjourney","name":"Midjourney","provider":"Midjourney Inc.","category":"creative","description":"A leading AI image generation tool that creates high-quality, artistic images from text prompts, known for its distinctive aesthetic style and photorealistic output.","url":"https://www.midjourney.com","pricing":"Basic $10/mo, Standard $30/mo, Pro $60/mo, Mega $120/mo","launched":2022}
+{"id":"gpt-image","name":"GPT Image","provider":"OpenAI","category":"creative","description":"OpenAI's natively multimodal image generation and editing models, built into ChatGPT and served through the Image and Responses APIs. GPT-Image-2.5 Sunburst is the flagship for precise edits and GPT-Image-2.5 Flare the fast everyday tier; they succeed DALL-E 3, which OpenAI removed from the API on May 12, 2026.","url":"https://developers.openai.com/api/docs/models/gpt-image-2.5-sunburst","pricing":"Included in ChatGPT plans; API GPT-Image-2.5 Sunburst at $5 per 1M text input tokens, $8 per 1M image input tokens, and $30 per 1M image output tokens","launched":2025}
+{"id":"suno","name":"Suno","provider":"Suno Inc.","category":"creative","description":"An AI music generation platform that creates full songs with vocals, instruments, and lyrics from text prompts or custom lyrics input.","url":"https://suno.com","pricing":"Free tier, Pro $10/mo, Premier $30/mo","launched":2023}
+{"id":"langchain","name":"LangChain","provider":"LangChain Inc.","category":"frameworks","description":"An open-source framework for building LLM-powered applications with chains, agents, retrieval-augmented generation, and memory, supporting multiple model providers.","url":"https://www.langchain.com","pricing":"Open source (LangSmith platform has paid tiers)","launched":2022}
+{"id":"crewai","name":"CrewAI","provider":"CrewAI Inc.","category":"frameworks","description":"A framework for orchestrating role-playing AI agents that collaborate on complex tasks, allowing developers to define agent roles, goals, and delegation patterns.","url":"https://www.crewai.com","pricing":"Open source, Enterprise cloud plans available","launched":2024}
+{"id":"microsoft-agent-framework","name":"Microsoft Agent Framework","provider":"Microsoft","category":"frameworks","description":"Microsoft's open-source framework for building agents and graph-based multi-agent workflows in Python and .NET, with a Go SDK in a separate repo. Microsoft calls it the enterprise-ready successor to AutoGen and Semantic Kernel, both of which ship migration guides to it; 1.0 shipped in April 2026 and AutoGen is now in maintenance mode.","url":"https://learn.microsoft.com/en-us/agent-framework/","pricing":"Open source (MIT)","launched":2025}
+{"id":"openai-agents-api","name":"OpenAI Agents API","provider":"OpenAI","category":"frameworks","description":"A hosted API, in public beta since September 10, 2026, that runs cloud agents on the same open-source harness behind Codex. One call sets the task, model, tools, MCP servers, and environment; OpenAI handles context compaction, tool search, and parallel subagents, with sandboxes hosted by OpenAI, on your own infrastructure, or with partners such as Cloudflare, Daytona, E2B, Modal, and Vercel. It supersedes the Assistants API, which OpenAI sunset on August 26, 2026.","url":"https://openai.com/index/introducing-the-agents-api/","pricing":"No additional fee for the API; pay for model tokens and tools at standard API rates","launched":2026}
+{"id":"claude-mcp","name":"Model Context Protocol (MCP)","provider":"Agentic AI Foundation (created by Anthropic)","category":"frameworks","description":"An open protocol that standardizes how AI applications connect to external data sources and tools, enabling plug-and-play integrations for any MCP-compatible client. Created at Anthropic and now hosted as a project of the Linux Foundation's Agentic AI Foundation; the current spec revision is 2026-07-28.","url":"https://modelcontextprotocol.io","pricing":"Open source protocol","launched":2024}

2026-10-12/ai-hardware.jsonl ADDED Viewed

	@@ -0,0 +1,26 @@

+{"id":"rubin","name":"NVIDIA Rubin (Vera Rubin)","manufacturer":"NVIDIA","family":"Rubin","process":"Not disclosed by NVIDIA","released":"2026-08","memoryGB":288,"memoryBandwidthTBs":19.2,"fp16TFLOPS":4000,"fp8TFLOPS":17500,"fp4TFLOPS":35000,"tdpWatts":null,"interconnect":"NVLink 6 (3 TB/s per GPU) + Vera CPU NVLink-C2C (1.8 TB/s)","listPriceUSD":null,"availability":"Cloud and rack-scale (Vera Rubin NVL72: 72 Rubin GPUs + 36 Vera CPUs); early racks at CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure, Nebius","notes":"Blackwell successor. 288GB HBM4 per GPU. NVIDIA began production shipments in August 2026. FLOPS are NVIDIA dense figures (FP8 is FP8/FP6 training, FP4 is NVFP4 training); the 50 PFLOPS NVFP4 inference headline is a sparse figure and is not used here.","url":"https://www.nvidia.com/en-us/data-center/vera-rubin-nvl72/"}
+{"id":"b300","name":"NVIDIA B300 (Blackwell Ultra)","manufacturer":"NVIDIA","family":"Blackwell Ultra","process":"TSMC 4NP","released":"2025","memoryGB":288,"memoryBandwidthTBs":8,"fp16TFLOPS":2500,"fp8TFLOPS":5000,"fp4TFLOPS":15000,"tdpWatts":1400,"interconnect":"NVLink 5 (1.8 TB/s)","listPriceUSD":null,"availability":"Cloud and rack-scale (GB300 NVL72: 72 Blackwell Ultra GPUs + 36 Grace CPUs), listed by NVIDIA as available now","notes":"Mid-cycle Blackwell refresh: 208 billion transistors across two dies, with NVIDIA quoting 1.5x the HBM3e capacity and 1.5x the dense NVFP4 rate of Blackwell. Dense figures shown; NVIDIA rates the NVL72 rack at 1,440 PFLOPS NVFP4 with sparsity and 1.1 exaFLOPS dense.","url":"https://www.nvidia.com/en-us/data-center/gb300-nvl72/"}
+{"id":"gb200","name":"NVIDIA GB200 (Blackwell)","manufacturer":"NVIDIA","family":"Blackwell","process":"TSMC 4NP","released":"2025","memoryGB":372,"memoryBandwidthTBs":16,"fp16TFLOPS":5000,"fp8TFLOPS":10000,"fp4TFLOPS":20000,"tdpWatts":2700,"interconnect":"NVLink 5 (1.8 TB/s) + Grace CPU C2C (900 GB/s)","listPriceUSD":null,"availability":"Cloud only (CoreWeave, AWS, Azure, GCP, Oracle, Lambda)","notes":"Blackwell flagship: dual B200 + Grace CPU. FP4 native. The chip every major lab is buying for 2025-2026 frontier training.","url":"https://www.nvidia.com/en-us/data-center/gb200-nvl72/"}
+{"id":"b200","name":"NVIDIA B200","manufacturer":"NVIDIA","family":"Blackwell","process":"TSMC 4NP","released":"2025","memoryGB":180,"memoryBandwidthTBs":8,"fp16TFLOPS":2500,"fp8TFLOPS":5000,"fp4TFLOPS":10000,"tdpWatts":1000,"interconnect":"NVLink 5 (1.8 TB/s)","listPriceUSD":null,"availability":"Cloud only","notes":"Single-die Blackwell. 180GB HBM3e per GPU (1,440GB across an 8-GPU DGX B200). Inference workhorse for serving 100B+ MoE models on a single GPU.","url":"https://www.nvidia.com/en-us/data-center/blackwell-architecture/"}
+{"id":"h200","name":"NVIDIA H200","manufacturer":"NVIDIA","family":"Hopper","process":"TSMC 4N","released":"2024","memoryGB":141,"memoryBandwidthTBs":4.8,"fp16TFLOPS":989,"fp8TFLOPS":1979,"fp4TFLOPS":null,"tdpWatts":700,"interconnect":"NVLink 4 (900 GB/s)","listPriceUSD":null,"availability":"Cloud and direct sale","notes":"Hopper refresh with HBM3e. The sweet spot for production inference in 2025. Extra memory means 70B-class models fit single-GPU at FP8.","url":"https://www.nvidia.com/en-us/data-center/h200/"}
+{"id":"h100","name":"NVIDIA H100","manufacturer":"NVIDIA","family":"Hopper","process":"TSMC 4N","released":"2022","memoryGB":80,"memoryBandwidthTBs":3.35,"fp16TFLOPS":989,"fp8TFLOPS":1979,"fp4TFLOPS":null,"tdpWatts":700,"interconnect":"NVLink 4 (900 GB/s)","listPriceUSD":30000,"availability":"Cloud and direct sale (~$30k retail, declining)","notes":"The chip that built the 2023-2024 LLM boom. Still the most-deployed AI GPU. FP8 native. 80GB HBM3.","url":"https://www.nvidia.com/en-us/data-center/h100/"}
+{"id":"a100","name":"NVIDIA A100 80GB","manufacturer":"NVIDIA","family":"Ampere","process":"TSMC 7nm","released":"2020","memoryGB":80,"memoryBandwidthTBs":2,"fp16TFLOPS":312,"fp8TFLOPS":null,"fp4TFLOPS":null,"tdpWatts":400,"interconnect":"NVLink 3 (600 GB/s)","listPriceUSD":18000,"availability":"Cloud and used market (~$18k retail, declining)","notes":"Workhorse of the late-2010s ML wave. Still widely used for training smaller models and inference where H100/H200 supply is tight.","url":"https://www.nvidia.com/en-us/data-center/a100/"}
+{"id":"rtx-pro-6000-blackwell","name":"NVIDIA RTX PRO 6000 Blackwell","manufacturer":"NVIDIA","family":"Blackwell","process":"TSMC 4NP","released":"2025","memoryGB":96,"memoryBandwidthTBs":1.79,"fp16TFLOPS":360,"fp8TFLOPS":720,"fp4TFLOPS":1440,"tdpWatts":600,"interconnect":"PCIe 5.0","listPriceUSD":8500,"availability":"Direct sale (workstation card)","notes":"Workstation Blackwell. 96GB GDDR7. Good fit for on-prem agent dev environments where multi-H100 rentals are overkill.","url":"https://www.nvidia.com/en-us/products/workstations/professional-desktop-gpus/rtx-pro-6000-blackwell/"}
+{"id":"rtx-4090","name":"NVIDIA RTX 4090","manufacturer":"NVIDIA","family":"Ada Lovelace","process":"TSMC 4N","released":"2022","memoryGB":24,"memoryBandwidthTBs":1.01,"fp16TFLOPS":165,"fp8TFLOPS":660,"fp4TFLOPS":null,"tdpWatts":450,"interconnect":"PCIe 4.0","listPriceUSD":1600,"availability":"Consumer retail","notes":"Best price/perf for local agent dev. 24GB VRAM fits Q4-quantized 70B models. The default consumer-tier choice for self-hosted Ollama/llama.cpp.","url":"https://www.nvidia.com/en-us/geforce/graphics-cards/40-series/rtx-4090/"}
+{"id":"rtx-5090","name":"NVIDIA RTX 5090","manufacturer":"NVIDIA","family":"Blackwell","process":"TSMC 4NP","released":"2025","memoryGB":32,"memoryBandwidthTBs":1.79,"fp16TFLOPS":209,"fp8TFLOPS":838,"fp4TFLOPS":3352,"tdpWatts":575,"interconnect":"PCIe 5.0","listPriceUSD":2000,"availability":"Consumer retail","notes":"Consumer Blackwell flagship. 32GB GDDR7. FP4 native. Strong pick for local agent dev with current-frontier features.","url":"https://www.nvidia.com/en-us/geforce/graphics-cards/50-series/rtx-5090/"}
+{"id":"mi455x","name":"AMD MI455X","manufacturer":"AMD","family":"Instinct","process":"Not disclosed by AMD","released":"2026-07","memoryGB":432,"memoryBandwidthTBs":23.3,"fp16TFLOPS":5030,"fp8TFLOPS":20130,"fp4TFLOPS":40260,"tdpWatts":null,"interconnect":"UALink over Ethernet (3.6 TB/s bidirectional scale-up per GPU) plus 2,400 Gb/s scale-out","listPriceUSD":null,"availability":"Rack-scale (AMD Helios: 72 MI455X GPUs and 18 6th Gen EPYC CPUs per rack) via Bull, HPE, Lenovo and Supermicro","notes":"AMD's MI400-series flagship, launched at Advancing AI on July 23, 2026 with Helios racks already in production. 432GB HBM4. Per-GPU figures are MXFP4, MXFP8 and dense BF16 peaks; AMD publishes the rack numbers of 2.9 exaflops FP4, 1.4 exaflops FP8, 31TB HBM4 and 1.7 PB/s across 72 GPUs. OpenAI expects to bring Helios online starting in Q4 2026.","url":"https://www.amd.com/en/products/accelerators/instinct/mi400/mi455x.html"}
+{"id":"mi355x","name":"AMD MI355X","manufacturer":"AMD","family":"Instinct","process":"TSMC N3 / N6","released":"2025-06","memoryGB":288,"memoryBandwidthTBs":8,"fp16TFLOPS":2500,"fp8TFLOPS":5000,"fp4TFLOPS":10100,"tdpWatts":1400,"interconnect":"Infinity Fabric (7x 153 GB/s scale-up + 1x 128 GB/s scale-out)","listPriceUSD":null,"availability":"Cloud (Oracle Cloud Infrastructure, Crusoe; DigitalOcean spot)","notes":"AMD's MI350-series flagship. 288GB HBM3e with native MXFP4 and MXFP6 support. 1400W typical board power. The step between MI325X and the Helios-rack MI400 series.","url":"https://www.amd.com/en/products/accelerators/instinct/mi350/mi355x.html"}
+{"id":"mi350p","name":"AMD Instinct MI350P","manufacturer":"AMD","family":"Instinct","process":"TSMC N3 / N6","released":"2026-05","memoryGB":144,"memoryBandwidthTBs":4,"fp16TFLOPS":1150,"fp8TFLOPS":2300,"fp4TFLOPS":4600,"tdpWatts":600,"interconnect":"PCIe 5.0 x16 (no GPU-to-GPU Infinity Fabric links)","listPriceUSD":null,"availability":"Direct sale and OEM servers (dual-slot, full-length air-cooled PCIe card; announced May 2026)","notes":"CDNA 4 in a standard PCIe slot: 128 compute units, roughly half an MI350X, with 144GB HBM3e and a configurable 450W power mode. Peak dense figures shown; AMD doubles its 8-bit and 16-bit ratings with structured sparsity. Built to drop inference capacity into existing air-cooled enterprise servers instead of liquid-cooled racks.","url":"https://www.amd.com/en/products/accelerators/instinct/mi350/mi350p.html"}
+{"id":"mi325x","name":"AMD MI325X","manufacturer":"AMD","family":"Instinct","process":"TSMC N5 / N6","released":"2024-Q4","memoryGB":256,"memoryBandwidthTBs":6,"fp16TFLOPS":1300,"fp8TFLOPS":2600,"fp4TFLOPS":null,"tdpWatts":1000,"interconnect":"Infinity Fabric (8x 153 GB/s)","listPriceUSD":null,"availability":"Cloud (Microsoft Azure, Oracle), direct sale","notes":"AMD's answer to H200. 256GB HBM3e (highest in the catalog at single-die). Strong inference value when supply is constrained on NVIDIA.","url":"https://www.amd.com/en/products/accelerators/instinct/mi300/mi325x.html"}
+{"id":"mi300x","name":"AMD MI300X","manufacturer":"AMD","family":"Instinct","process":"TSMC N5 / N6","released":"2023-Q4","memoryGB":192,"memoryBandwidthTBs":5.3,"fp16TFLOPS":1300,"fp8TFLOPS":2600,"fp4TFLOPS":null,"tdpWatts":750,"interconnect":"Infinity Fabric","listPriceUSD":15000,"availability":"Cloud (Vultr, RunPod, Microsoft Azure), direct sale (~$15k)","notes":"Wider availability than NVIDIA flagship; cheaper to rent. ROCm software stack matures; vLLM and PyTorch first-class on MI300X in 2026.","url":"https://www.amd.com/en/products/accelerators/instinct/mi300/mi300x.html"}
+{"id":"tpu7x-ironwood","name":"Google TPU7x (Ironwood)","manufacturer":"Google","family":"TPU","process":"Not disclosed by Google","released":"2026-03","memoryGB":192,"memoryBandwidthTBs":7.38,"fp16TFLOPS":2307,"fp8TFLOPS":4614,"fp4TFLOPS":null,"tdpWatts":null,"interconnect":"ICI (1.2 TB/s bidirectional per chip), 9,216-chip pods","listPriceUSD":null,"availability":"Google Cloud only (TPU7x VMs; preview November 2025, GA March 2026)","notes":"Seventh-generation TPU and the first in the Ironwood family. 192 GiB HBM per chip, six times Trillium. Google positions it for large-scale training and inference of LLMs, MoE, and diffusion models.","url":"https://docs.cloud.google.com/tpu/docs/tpu7x"}
+{"id":"tpu-v6e-trillium","name":"Google TPU v6e (Trillium)","manufacturer":"Google","family":"TPU","process":"Not disclosed by Google","released":"2024-12","memoryGB":32,"memoryBandwidthTBs":1.64,"fp16TFLOPS":918,"fp8TFLOPS":918,"fp4TFLOPS":null,"tdpWatts":null,"interconnect":"ICI (800 GB/s bidirectional per chip), 256-chip pods","listPriceUSD":null,"availability":"Google Cloud only (GA December 2024)","notes":"Sixth-generation TPU. Twice the per-chip BF16 compute of v5p with a third of the memory; Google used it to train Gemini 2.0.","url":"https://docs.cloud.google.com/compute/docs/tpus/tpu-machines"}
+{"id":"tpu-v5p","name":"Google TPU v5p","manufacturer":"Google","family":"TPU","process":"TSMC 3nm","released":"2023-12","memoryGB":95,"memoryBandwidthTBs":2.77,"fp16TFLOPS":459,"fp8TFLOPS":459,"fp4TFLOPS":null,"tdpWatts":700,"interconnect":"ICI (Inter-Chip Interconnect, 1.2 TB/s bidirectional per chip)","listPriceUSD":null,"availability":"Google Cloud only (Vertex AI, GCE)","notes":"Google's previous-generation training TPU, now behind Trillium and Ironwood. Used for Gemini training. Only available on GCP; pod-scale (8960 chips) is a different beast than per-chip rentals.","url":"https://cloud.google.com/tpu/docs/v5p"}
+{"id":"tpu-v5e","name":"Google TPU v5e","manufacturer":"Google","family":"TPU","process":"TSMC 5nm","released":"2023-08","memoryGB":16,"memoryBandwidthTBs":0.8,"fp16TFLOPS":197,"fp8TFLOPS":393,"fp4TFLOPS":null,"tdpWatts":170,"interconnect":"ICI (400 GB/s bidirectional per chip)","listPriceUSD":null,"availability":"Google Cloud only","notes":"TPU inference tier. The 8-bit figure is Google's INT8 rating (393 TOPS). Cheaper than v5p; 256-chip pods. Best fit for Google Cloud customers serving Gemini-style inference at scale.","url":"https://cloud.google.com/tpu/docs/v5e"}
+{"id":"trainium-2","name":"AWS Trainium 2","manufacturer":"AWS","family":"Trainium","process":"TSMC 5nm","released":"2024-12","memoryGB":96,"memoryBandwidthTBs":2.9,"fp16TFLOPS":667,"fp8TFLOPS":1299,"fp4TFLOPS":null,"tdpWatts":500,"interconnect":"NeuronLink-v3 (1.28 TB/s per chip)","listPriceUSD":null,"availability":"AWS only (Trn2 instances)","notes":"AWS custom training/inference silicon. Dense figures shown; AWS rates it at 2,563 TFLOPS with sparsity. Anthropic uses these for Claude training. Cheaper TCO on AWS than NVIDIA flagship for committed workloads.","url":"https://aws.amazon.com/ai/machine-learning/trainium/"}
+{"id":"trainium-3","name":"AWS Trainium3","manufacturer":"AWS","family":"Trainium","process":"TSMC 3nm","released":"2025-12","memoryGB":144,"memoryBandwidthTBs":4.9,"fp16TFLOPS":671,"fp8TFLOPS":2517,"fp4TFLOPS":2517,"tdpWatts":null,"interconnect":"NeuronLink-v4 (2.56 TB/s per chip)","listPriceUSD":null,"availability":"AWS only (Trn3 instances)","notes":"Third-generation AWS training and inference silicon. 144 GB HBM3e, 4.9 TB/s bandwidth. 8-bit and 4-bit figures are MXFP8 and MXFP4 per AWS; BF16 is roughly flat with Trainium 2. Roughly 30-50% lower cost per hour than H100/H200 on committed AWS workloads.","url":"https://aws.amazon.com/ai/machine-learning/trainium/"}
+{"id":"inferentia-2","name":"AWS Inferentia 2","manufacturer":"AWS","family":"Inferentia","process":"TSMC 5nm","released":"2023-04","memoryGB":32,"memoryBandwidthTBs":0.82,"fp16TFLOPS":190,"fp8TFLOPS":null,"fp4TFLOPS":null,"tdpWatts":null,"interconnect":"NeuronLink","listPriceUSD":null,"availability":"AWS only (Inf2 instances)","notes":"AWS inference silicon. Lower compute than Trainium 2 but cheaper per-token for high-volume serving workloads on AWS.","url":"https://aws.amazon.com/ai/machine-learning/inferentia/"}
+{"id":"apple-m4-max","name":"Apple M4 Max","manufacturer":"Apple","family":"Apple Silicon","process":"TSMC 3nm","released":"2024-10","memoryGB":128,"memoryBandwidthTBs":0.546,"fp16TFLOPS":38,"fp8TFLOPS":null,"fp4TFLOPS":null,"tdpWatts":60,"interconnect":"On-die (unified memory)","listPriceUSD":4699,"availability":"Prior generation: Apple ships M5 Max (up to 128GB, 614 GB/s) in the MacBook Pro from March 11, 2026 and in the Mac Studio from September 22, 2026, where M5 Ultra reaches 512GB of unified memory at 1.2 TB/s","notes":"128GB unified memory means a 2024-2025 Mac Studio or MacBook Pro runs 70B-class models at Q4 quantization with usable speed. The dark-horse local-inference platform. Apple publishes memory and bandwidth for Apple Silicon but no TFLOPS rating for the M5 generation, so no M5 row is listed here.","url":"https://www.apple.com/newsroom/2024/10/apple-introduces-m4-pro-and-m4-max/"}
+{"id":"cerebras-wse-3-turbo","name":"Cerebras WSE-3 Turbo","manufacturer":"Cerebras","family":"WSE","process":"Not disclosed by Cerebras","released":"2026-08","memoryGB":44,"memoryBandwidthTBs":43200,"fp16TFLOPS":250000,"fp8TFLOPS":null,"fp4TFLOPS":null,"tdpWatts":null,"interconnect":"On-wafer fabric (53.5 PB/s), 2.4 Tb/s off-wafer I/O","listPriceUSD":null,"availability":"Cerebras Cloud and CS-4 systems (three WSE-3T wafers per CS-4 rack)","notes":"Announced August 18, 2026. Same 46,225 mm^2 wafer, 4 trillion transistors and 900,000 cores as WSE-3, but Cerebras doubles AI compute to 250 PFLOPS and memory bandwidth to 43.2 PB/s (shown here as 43,200 TB/s). A CS-4 rack holds three wafers for 750 PFLOPS. Compute figure is on the same basis as the 125 PFLOPS FP16 rating of WSE-3.","url":"https://www.cerebras.ai/chip"}
+{"id":"cerebras-wse-3","name":"Cerebras WSE-3","manufacturer":"Cerebras","family":"WSE","process":"TSMC 5nm","released":"2024","memoryGB":44,"memoryBandwidthTBs":21000,"fp16TFLOPS":125000,"fp8TFLOPS":null,"fp4TFLOPS":null,"tdpWatts":23000,"interconnect":"On-wafer (no interconnect needed)","listPriceUSD":null,"availability":"Cerebras Cloud only; superseded by WSE-3 Turbo in August 2026","notes":"Single 46,225 mm^2 wafer. Holds entire activations on-chip; eliminates many distributed-training pains. Cerebras rates aggregate memory bandwidth at 21 PB/s (shown here as 21,000 TB/s).","url":"https://cerebras.ai/product-chip/"}
+{"id":"groq-lpu","name":"Groq LPU","manufacturer":"Groq","family":"LPU","process":"GlobalFoundries 14nm (gen 1)","released":"2023","memoryGB":0.23,"memoryBandwidthTBs":80,"fp16TFLOPS":188,"fp8TFLOPS":null,"fp4TFLOPS":null,"tdpWatts":350,"interconnect":"GroqLink","listPriceUSD":null,"availability":"Groq Cloud only","notes":"Deterministic single-thread inference silicon. SRAM-only memory model means small per-chip capacity but massive bandwidth. Behind Groq's 700+ tokens/sec Llama 4 Scout serving.","url":"https://groq.com/lpu-language-processing-unit/"}

2026-10-12/ai-lawsuits.jsonl ADDED Viewed

	@@ -0,0 +1,51 @@

+{"id":"nyt-v-openai","name":"The New York Times Co. v. Microsoft Corp. & OpenAI","plaintiff":"The New York Times Company","defendants":["Microsoft Corporation","OpenAI, Inc.","OpenAI LP","OpenAI GP"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:23-cv-11195 (part of MDL No. 3143)","filed":"2023-12-27","status":"active","stage":"summary-judgment","claims":["copyright-infringement","dmca-violation","unfair-competition"],"summary":"NYT alleges OpenAI and Microsoft trained GPT-class models on millions of NYT articles without license and that ChatGPT outputs reproduce articles verbatim. Since April 2025 it has been coordinated in MDL No. 3143 (In re OpenAI, Inc. Copyright Infringement Litigation) before Judge Sidney Stein; an April 4, 2025 opinion dismissed some DMCA claims while the core copyright claims proceeded, and in January 2026 Judge Stein affirmed orders requiring OpenAI to produce a 20 million ChatGPT log sample. The Times, OpenAI, and Microsoft cross-moved for summary judgment on September 4, 2026, days after the U.S. Justice Department filed a statement of interest arguing that training on copyrighted works can qualify as fair use.","sources":["https://www.courtlistener.com/docket/68117049/the-new-york-times-company-v-microsoft-corporation/","https://www.jpml.uscourts.gov/sites/jpml/files/MDL-3143-Transfer_Order-3-25.pdf","https://www.nysd.uscourts.gov/sites/default/files/2025-04/yf%2023cv11195%20OpenAI%20MTD%20opinion%20april%204%202025.pdf","https://www.axios.com/2026/09/08/nyt-openai-microsoft-copyright-lawsuit"]}
+{"id":"daily-news-v-openai","name":"Daily News, LP et al. v. Microsoft Corp. & OpenAI","plaintiff":"Daily News LP, Chicago Tribune, Denver Post, Orlando Sentinel, Sun Sentinel, San Jose Mercury News, Orange County Register, St. Paul Pioneer Press (eight Alden Global Capital papers)","defendants":["Microsoft Corporation","OpenAI, Inc.","OpenAI subsidiaries"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:24-cv-03285 (part of MDL No. 3143)","filed":"2024-04-30","status":"consolidated","stage":"summary-judgment","claims":["copyright-infringement","dmca-violation","trademark-infringement"],"summary":"Eight Alden-owned newspapers allege the same training-data and output copying as the NYT case and are coordinated with it in MDL No. 3143. The April 4, 2025 opinion let DMCA 1202(b)(1) claims against OpenAI and trademark dilution claims proceed while dismissing the DMCA claims against Microsoft and the misappropriation claims. The papers joined the Times in a July 2026 motion for evidentiary sanctions over ChatGPT log discovery and moved for summary judgment on September 4, 2026.","sources":["https://www.courtlistener.com/docket/68484432/daily-news-lp-v-microsoft-corporation/","https://www.nysd.uscourts.gov/sites/default/files/2025-04/yf%2023cv11195%20OpenAI%20MTD%20opinion%20april%204%202025.pdf","https://techcrunch.com/2026/07/09/new-york-times-says-openai-hid-evidence-in-chatgpt-copyright-trial/"]}
+{"id":"cir-v-openai","name":"Center for Investigative Reporting v. OpenAI & Microsoft","plaintiff":"Center for Investigative Reporting (Mother Jones, Reveal)","defendants":["OpenAI, Inc.","Microsoft Corporation"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:24-cv-04872 (part of MDL No. 3143)","filed":"2024-06-27","status":"active","stage":"summary-judgment","claims":["copyright-infringement","dmca-violation"],"summary":"Investigative-journalism nonprofit alleges OpenAI scraped CIR's reporting from Mother Jones and Reveal for training, and that ChatGPT regurgitates articles without attribution. Transferred into MDL No. 3143 in April 2025; the April 4, 2025 opinion dismissed the misappropriation and abridgment claims with prejudice but let the DMCA 1202(b)(1) claim against OpenAI proceed. CIR was among the news plaintiffs that moved for summary judgment on September 4, 2026.","sources":["https://www.courtlistener.com/docket/68892274/the-center-for-investigative-reporting-inc-v-openai-inc/","https://www.nysd.uscourts.gov/sites/default/files/2025-04/yf%2023cv11195%20OpenAI%20MTD%20opinion%20april%204%202025.pdf"]}
+{"id":"raw-story-v-openai","name":"Raw Story Media v. OpenAI","plaintiff":"Raw Story Media; AlterNet Media","defendants":["OpenAI, Inc."],"jurisdiction":"US (S.D.N.Y.; 2d Cir.)","court":"United States District Court for the Southern District of New York; U.S. Court of Appeals for the Second Circuit","caseNumber":"1:24-cv-01514 (2d Cir. No. 25-1756)","filed":"2024-02-28","status":"dismissed","stage":"appeal","claims":["dmca-violation"],"summary":"DMCA-only theory (CMI removal) without a copyright claim. Judge McMahon dismissed for lack of standing in November 2024 and denied leave to amend on April 3, 2025; Judge Stein, as MDL judge, denied reconsideration on June 18, 2025. The Second Circuit heard the plaintiffs' appeal on March 18, 2026 and had not ruled as of September 2026.","sources":["https://www.courtlistener.com/docket/68290709/raw-story-media-inc-v-openai-inc/","https://law.justia.com/cases/federal/district-courts/new-york/nysdce/1:2024cv01514/616533/178/","https://courthousenews.com/digital-news-outlets-urge-second-circuit-to-revive-chatgpt-copyright-claims/"]}
+{"id":"ziff-davis-v-openai","name":"Ziff Davis v. OpenAI","plaintiff":"Ziff Davis, Inc. and affiliated publishers","defendants":["OpenAI, Inc. and affiliated entities"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York (transferred from D. Del.)","caseNumber":"1:25-cv-04315 (part of MDL No. 3143)","filed":"2025-04-24","status":"active","stage":"summary-judgment","claims":["copyright-infringement","dmca-violation"],"summary":"The digital publisher alleges OpenAI copied content from its sites for training and outputs. Filed in Delaware and transferred into MDL No. 3143; a December 15, 2025 opinion dismissed some claims, including a DMCA anti-circumvention claim based on robots.txt, while other claims proceeded. Ziff Davis moved for summary judgment with the other news plaintiffs on September 4, 2026.","sources":["https://law.justia.com/cases/federal/district-courts/new-york/nysdce/1:2025cv04315/643043/300/","https://www.axios.com/2026/09/08/nyt-openai-microsoft-copyright-lawsuit"]}
+{"id":"richner-v-openai","name":"Richner Communications, Inc. et al. v. Microsoft & OpenAI","plaintiff":"Richner Communications, Inc. and other local newspaper publishers","defendants":["Microsoft Corporation","OpenAI, Inc. and affiliated entities"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:26-cv-05320","filed":"2026-06-24","status":"active","stage":"complaint","claims":["copyright-infringement","dmca-violation"],"summary":"A group of local newspaper publishers representing hundreds of outlets alleges OpenAI and Microsoft copied their articles for training and removed copyright management information. Marked related to MDL No. 3143, where it was stayed by stipulation in July 2026 pending the MDL summary judgment rulings.","sources":["https://www.courthousenews.com/wp-content/uploads/2026/06/Richner-publishers-openAI-complaint.pdf","https://www.arkansasonline.com/news/2026/jun/25/arkansas-democrat-gazette-wehco-join-lawsuit/"]}
+{"id":"seattle-times-newsday-v-openai","name":"The Seattle Times Co. & Newsday LLC v. OpenAI & Microsoft","plaintiff":"The Seattle Times Company; Newsday LLC","defendants":["OpenAI, Inc. and affiliated entities","Microsoft Corporation"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:26-cv-07644","filed":"2026-09-04","status":"active","stage":"complaint","claims":["copyright-infringement","dmca-violation","trademark-infringement"],"summary":"Two regional newspapers allege OpenAI and Microsoft scraped their journalism, including paywalled articles, into training datasets, removed copyright management information, and reproduce passages in outputs. The complaint adds federal and state trademark dilution claims over hallucinated content attributed to the papers and seeks destruction of models and datasets that incorporate their works.","sources":["https://www.medianama.com/wp-content/uploads/2026/09/gov.uscourts.nysd_.672142.1.0.pdf","https://www.theverge.com/ai-artificial-intelligence/990932/seattle-times-newsday-lawsuit-openai-microsoft"]}
+{"id":"wikihow-v-openai","name":"wikiHow v. OpenAI","plaintiff":"wikiHow","defendants":["OpenAI, Inc. and affiliated entities"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:26-cv-07171","filed":"2026-08-21","status":"active","stage":"complaint","claims":["copyright-infringement","dmca-violation"],"summary":"The how-to publisher alleges OpenAI copied more than 11,000 of its articles for AI training, bringing direct and vicarious infringement and DMCA 1202(b)(1) claims. Stayed by joint stipulation in September 2026 pending summary judgment rulings in MDL No. 3143.","sources":["https://www.reuters.com/legal/legalindustry/wikihow-sues-openai-copyright-infringement-over-ai-training-2026-08-24/"]}
+{"id":"newscorp-v-perplexity","name":"Dow Jones & NYP Holdings v. Perplexity AI","plaintiff":"Dow Jones & Company; NYP Holdings (News Corp)","defendants":["Perplexity AI, Inc."],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:24-cv-07984","filed":"2024-10-21","status":"active","stage":"discovery","claims":["copyright-infringement","unfair-competition","trademark-infringement"],"summary":"WSJ and NY Post owners allege Perplexity's answer engine copies and republishes their reporting verbatim and falsely attributes its summaries. Judge Failla denied Perplexity's motion to dismiss or transfer in full on August 21, 2025, and 2026 discovery orders have compelled production of additional Perplexity logs and searches. A pretrial conference is set for December 14, 2026.","sources":["https://www.courtlistener.com/docket/69280523/dow-jones-company-inc-v-perplexity-ai-inc/","https://cases.justia.com/federal/district-courts/new-york/nysdce/1:2024cv07984/630270/65/0.pdf"]}
+{"id":"britannica-v-perplexity","name":"Encyclopaedia Britannica & Merriam-Webster v. Perplexity AI","plaintiff":"Encyclopaedia Britannica, Inc.; Merriam-Webster, Inc.","defendants":["Perplexity AI, Inc."],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:25-cv-07546","filed":"2025-09-10","status":"active","stage":"discovery","claims":["copyright-infringement","trademark-infringement"],"summary":"The reference publishers allege Perplexity's answer engine copies their entries and definitions and diverts their traffic, and that it attaches their marks to hallucinated or incomplete content. Discovery is under way while Perplexity's partial motion to dismiss the output-based claim remains pending.","sources":["https://storage.courtlistener.com/recap/gov.uscourts.nysd.649196/gov.uscourts.nysd.649196.1.0.pdf","https://www.reuters.com/legal/litigation/encyclopedia-britannica-sues-perplexity-over-ai-answer-engine-2025-09-11/"]}
+{"id":"nyt-v-perplexity","name":"The New York Times Co. v. Perplexity AI","plaintiff":"The New York Times Company","defendants":["Perplexity AI, Inc."],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:25-cv-10106","filed":"2025-12-05","status":"active","stage":"motion-to-dismiss","claims":["copyright-infringement","trademark-infringement"],"summary":"The Times alleges Perplexity copies and distributes its journalism, including paywalled content, through retrieval-augmented generation and attributes hallucinated text to the paper. Judge Preska accepted it as related to Chicago Tribune v. Perplexity (1:25-cv-10094); Perplexity's renewed partial motion to dismiss, filed April 2026, targets the output-based, secondary liability, and trademark counts and was pending as of September 2026.","sources":["https://www.courtlistener.com/docket/71997720/the-new-york-times-company-v-perplexity-ai-inc/","https://www.nytco.com/press/the-times-sues-perplexity-ai/","https://www.reuters.com/legal/litigation/new-york-times-sues-perplexity-ai-infringing-copyright-works-2025-12-05/"]}
+{"id":"cnn-v-perplexity","name":"Cable News Network, Inc. v. Perplexity AI","plaintiff":"Cable News Network, Inc.","defendants":["Perplexity AI, Inc."],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:26-cv-04427","filed":"2026-05-28","status":"active","stage":"complaint","claims":["copyright-infringement","trademark-infringement"],"summary":"CNN alleges Perplexity copies its reporting and violates its trademarks under the Lanham Act by implying an affiliation between the two companies that does not exist.","sources":["https://variety.com/wp-content/uploads/2026/05/CNN-v-Perplexity.pdf","https://www.cnn.com/2026/05/28/media/cnn-sues-perplexity-ai-copyright"]}
+{"id":"yomiuri-v-perplexity","name":"Yomiuri Shimbun v. Perplexity AI","plaintiff":"The Yomiuri Shimbun group companies","defendants":["Perplexity AI, Inc."],"jurisdiction":"Japan","court":"Tokyo District Court","caseNumber":null,"filed":"2025-08-07","status":"active","stage":"motion-to-dismiss","claims":["copyright-infringement"],"summary":"Copyright suit by Yomiuri Shimbun group companies seeking roughly 2.17 billion yen in damages from Perplexity. At the first oral hearing in May 2026, Perplexity asked the court to reject the suit.","sources":["https://www.yomiuri.co.jp/national/20260526-GYT1T00339/"]}
+{"id":"nikkei-asahi-v-perplexity","name":"Nikkei & Asahi Shimbun v. Perplexity AI","plaintiff":"Nikkei Inc.; The Asahi Shimbun Company","defendants":["Perplexity AI, Inc."],"jurisdiction":"Japan","court":"Tokyo District Court","caseNumber":null,"filed":"2025-08-26","status":"active","stage":"motion-to-dismiss","claims":["copyright-infringement","unfair-competition"],"summary":"The two publishers seek an injunction, deletion of stored articles, and 2.2 billion yen each in damages, alleging Perplexity ignored robots.txt, copied articles including paywalled Nikkei content, and displayed their names alongside inaccurate answers in violation of Japan's Unfair Competition Prevention Act.","sources":["https://www.nikkei.co.jp/nikkeiinfo/en/news/release_en_20250826_01.pdf","https://www.japantimes.co.jp/news/2025/08/26/japan/crime-legal/japan-newspapers-sue-ai-startup/"]}
+{"id":"advance-local-v-cohere","name":"Advance Local Media et al. v. Cohere","plaintiff":"Advance Local Media, Conde Nast, The Atlantic, Forbes, Guardian News & Media, Insider, Los Angeles Times, McClatchy, Newsday, Plain Dealer, Politico, The Republican, Toronto Star, Vox Media","defendants":["Cohere Inc."],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:25-cv-01305","filed":"2025-02-13","status":"active","stage":"discovery","claims":["copyright-infringement","trademark-infringement"],"summary":"Fourteen publishers allege Cohere copies their articles to train its Command models and serves verbatim copies, excerpts, and substitutive summaries through retrieval, and that hallucinated articles bear their marks. Judge McMahon denied Cohere's partial motion to dismiss on November 13, 2025; summary judgment motions are due December 18, 2026.","sources":["https://www.newsmediaalliance.org/wp-content/uploads/2025/11/Doc-59-Decision-and-Order-Denying-Partial-Motion-to-Dismiss.pdf","https://case-law.vlex.com/vid/advance-local-media-llc-1094877471"]}
+{"id":"penske-v-google","name":"Penske Media Corp. v. Google (AI Overviews)","plaintiff":"Penske Media Corporation and affiliated publications (Rolling Stone, Billboard, Variety, The Hollywood Reporter, Deadline, others)","defendants":["Google LLC","Alphabet Inc."],"jurisdiction":"US (D.D.C.)","court":"United States District Court for the District of Columbia","caseNumber":"1:25-cv-03192","filed":"2025-09-12","status":"active","stage":"motion-to-dismiss","claims":["antitrust"],"summary":"First major US publisher suit over Google's AI Overviews, alleging Google uses its search monopoly to force publishers to supply content for AI summaries that cut referral traffic. Judge Mehta held a hearing on Google's motion to dismiss on August 25, 2026, consolidated with Chegg v. Google, and had not ruled as of late August 2026.","sources":["https://techcrunch.com/2025/09/14/rolling-stone-owner-penske-media-sues-google-over-ai-summaries/"]}
+{"id":"thomson-reuters-v-ross","name":"Thomson Reuters Enterprise Centre GmbH v. ROSS Intelligence","plaintiff":"Thomson Reuters Enterprise Centre GmbH","defendants":["ROSS Intelligence Inc."],"jurisdiction":"US (D. Del.; 3d Cir.)","court":"United States District Court for the District of Delaware; U.S. Court of Appeals for the Third Circuit","caseNumber":"1:20-cv-00613 (3d Cir. No. 25-2153)","filed":"2020-05-06","status":"active","stage":"appeal","claims":["copyright-infringement"],"summary":"Thomson Reuters alleges ROSS copied Westlaw headnotes to build an AI legal research tool. On February 11, 2025 Judge Stephanos Bibas, sitting by designation, granted Thomson Reuters partial summary judgment on direct infringement for 2,243 headnotes and rejected ROSS's fair use defense, noting only non-generative AI was before him. He certified the originality and fair use questions for interlocutory appeal; the Third Circuit heard argument on June 11, 2026 and had not ruled as of September 2026.","sources":["https://www.hunton.com/assets/htmldocuments/Thomson-Reuters-v-ROSS.pdf","https://www.courtlistener.com/docket/17131648/thomson-reuters-enterprise-centre-gmbh-v-ross-intelligence-inc/","https://www.bakerbotts.com/thought-leadership/publications/2026/july/third-circuit-hears-oral-argument"]}
+{"id":"authors-guild-v-openai","name":"Authors Guild et al. v. OpenAI (consolidated)","plaintiff":"Authors Guild, John Grisham, George R.R. Martin, Jodi Picoult, David Baldacci, Michael Connelly, others","defendants":["OpenAI, Inc.","Microsoft Corporation"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:23-cv-08292 (part of MDL No. 3143)","filed":"2023-09-19","status":"consolidated","stage":"summary-judgment","claims":["copyright-infringement"],"summary":"Putative class action by novelists and the Authors Guild alleging OpenAI trained on copyrighted books. Now part of the book-author class track in MDL No. 3143 with the transferred Tremblay and Silverman cases; on October 27, 2025 Judge Stein denied OpenAI's motion to dismiss the consolidated complaint's output-based infringement claim. Summary judgment motions were filed by the September 4, 2026 deadline, and class certification briefing is scheduled to follow the summary judgment ruling.","sources":["https://authorsguild.org/news/ag-and-authors-file-class-action-suit-against-openai/","https://www.jpml.uscourts.gov/sites/jpml/files/MDL-3143-Transfer_Order-3-25.pdf"]}
+{"id":"silverman-v-openai","name":"Tremblay / Silverman v. OpenAI","plaintiff":"Paul Tremblay, Sarah Silverman, Christopher Golden, Richard Kadrey, and other authors","defendants":["OpenAI, Inc."],"jurisdiction":"US (N.D. Cal., transferred to S.D.N.Y.)","court":"United States District Court for the Southern District of New York (transferred from the Northern District of California)","caseNumber":"3:23-cv-03223 and 3:23-cv-03416 (N.D. Cal.); 1:25-cv-03482 and 1:25-cv-03483 (S.D.N.Y., MDL No. 3143)","filed":"2023-06-28","status":"consolidated","stage":"summary-judgment","claims":["copyright-infringement"],"summary":"Related author class actions alleging OpenAI trained on copyrighted books obtained from shadow libraries. In N.D. Cal. the vicarious infringement, DMCA, negligence, and unjust enrichment claims were dismissed in February 2024 and the UCL claim in July 2024, leaving direct infringement. The cases were transferred into MDL No. 3143 in April 2025 and proceed on its book-author class track.","sources":["https://www.courthousenews.com/wp-content/uploads/2024/07/ucl-claim-dismiss-openai-case.pdf","https://www.jpml.uscourts.gov/sites/jpml/files/MDL-3143-Transfer_Order-3-25.pdf"]}
+{"id":"kadrey-v-meta","name":"Kadrey v. Meta Platforms (Llama)","plaintiff":"Richard Kadrey, Sarah Silverman, Christopher Golden, and other authors","defendants":["Meta Platforms, Inc."],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California","caseNumber":"3:23-cv-03417","filed":"2023-07-07","status":"active","stage":"discovery","claims":["copyright-infringement"],"summary":"Authors allege Meta trained Llama on pirated books. On June 25, 2025 Judge Chhabria granted Meta partial summary judgment on fair use as to training for these plaintiffs, while stating the ruling does not mean Meta's training use is lawful and leaving the torrent distribution claim live. In 2026 he allowed a contributory infringement claim based on BitTorrent seeding and on July 8 denied the authors' request for an interlocutory appeal; summary judgment briefing on the remaining claims is set to begin in December 2026.","sources":["https://law.justia.com/cases/federal/district-courts/california/candce/3:2023cv03417/415175/598/","https://torrentfreak.com/judge-allows-bittorrent-seeding-claims-against-meta-despite-lawyers-lame-excuses/","https://chatgptiseatingtheworld.com/2026/07/08/judge-chhabria-denies-kadrey-book-authors-motion-to-seek-an-interlocutory-appeal-of-downloading-part-of-ruling-in-fair-use-decision/"]}
+{"id":"bartz-v-anthropic","name":"Bartz v. Anthropic","plaintiff":"Andrea Bartz and other authors (certified class)","defendants":["Anthropic PBC"],"jurisdiction":"US (N.D. Cal.; 9th Cir.)","court":"United States District Court for the Northern District of California","caseNumber":"4:24-cv-05417 (9th Cir. Nos. 26-5391, 26-5455)","filed":"2024-08-19","status":"settled","stage":"appeal","claims":["copyright-infringement"],"summary":"Authors alleged Anthropic trained Claude on books, including copies downloaded from pirate libraries. In June 2025 Judge Alsup held that training and print-to-digital conversion were fair use but that building a central library from pirated copies was not, then certified a class; the parties settled for $1.5 billion (about $3,000 per work). Judge Martinez-Olguin granted final approval and entered judgment on July 20, 2026, and appeals of the approval and fee order were filed in August 2026.","sources":["https://www.courtlistener.com/docket/69058235/bartz-v-anthropic-pbc/","https://authorsguild.org/news/court-grants-final-approval-anthropic-copyright-settlement/","https://www.anthropiccopyrightsettlement.com/"]}
+{"id":"opt-out-authors-v-anthropic","name":"Shakespeare et al. v. Anthropic (Bartz settlement opt-outs)","plaintiff":"Authors who opted out of the Bartz v. Anthropic class settlement","defendants":["Anthropic PBC","Dario Amodei","Benjamin Mann"],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California","caseNumber":"5:26-cv-05931","filed":"2026-06-17","status":"active","stage":"complaint","claims":["copyright-infringement"],"summary":"More than 100 authors who excluded themselves from the $1.5 billion Bartz settlement filed their own infringement suit over Anthropic's use of their books, also naming two Anthropic co-founders.","sources":["https://lunch.publishersmarketplace.com/2026/06/authors-who-opted-out-of-settlement-sue-anthropic-for-infringement/"]}
+{"id":"publishers-v-google-gemini","name":"Hachette Book Group et al. v. Google (Gemini)","plaintiff":"Hachette Book Group, Cengage, Elsevier, Scott Turow, and others (putative class)","defendants":["Google LLC"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:26-cv-05870","filed":"2026-07-10","status":"active","stage":"complaint","claims":["copyright-infringement"],"summary":"Book and academic publishers plus author Scott Turow allege Google copied copyrighted books to train its Gemini models. Several of the same publishers and Turow filed a parallel suit against Meta and Mark Zuckerberg in May 2026 (S.D.N.Y. 1:26-cv-03689).","sources":["https://publishingperspectives.com/2026/07/u-s-publishers-sue-google-alleging-massive-copyright-infringement-behind-its-gemini-ai-service/"]}
+{"id":"doe-v-github","name":"Doe et al. v. GitHub, OpenAI, Microsoft (Copilot)","plaintiff":"Anonymous open-source developers","defendants":["GitHub, Inc.","Microsoft Corporation","OpenAI, Inc."],"jurisdiction":"US (N.D. Cal.; 9th Cir.)","court":"United States District Court for the Northern District of California; U.S. Court of Appeals for the Ninth Circuit","caseNumber":"4:22-cv-06823 (9th Cir. No. 24-7700)","filed":"2022-11-03","status":"active","stage":"appeal","claims":["dmca-violation","breach-of-contract"],"summary":"Putative class action by open-source developers alleging GitHub Copilot reproduces their code without honoring license attribution requirements. Judge Tigar dismissed the DMCA 1202(b) claim with prejudice in June 2024 for lack of identicality, letting only breach-of-contract claims over open-source licenses proceed, then certified the identicality question for interlocutory appeal and stayed the case. The Ninth Circuit heard argument on February 11, 2026; no opinion had issued as of mid-September 2026.","sources":["https://storage.courtlistener.com/recap/gov.uscourts.cand.403220/gov.uscourts.cand.403220.253.0_1.pdf","https://docs.justia.com/cases/federal/district-courts/california/candce/4:2022cv06823/403220/282","https://githubcopilotlitigation.com"]}
+{"id":"andersen-v-stability","name":"Andersen v. Stability AI, Midjourney, DeviantArt, Runway","plaintiff":"Sarah Andersen, Kelly McKernan, Karla Ortiz, others (artist class)","defendants":["Stability AI Ltd.","Stability AI, Inc.","Midjourney, Inc.","DeviantArt, Inc.","Runway AI, Inc."],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California","caseNumber":"3:23-cv-00201","filed":"2023-01-13","status":"active","stage":"discovery","claims":["copyright-infringement","trademark-infringement"],"summary":"Class action by visual artists alleging diffusion-model training on LAION-scraped artwork without license. Runway was added in the amended complaint; in August 2024 Judge Orrick let the copyright claims against all four defendants and a Lanham Act claim against Midjourney proceed while dismissing the DMCA claims with prejudice (right of publicity claims had been dismissed in 2023). A June 2026 scheduling order set dispositive motions for June 2027 and a jury trial for September 20, 2027.","sources":["https://storage.courtlistener.com/recap/gov.uscourts.cand.407208/gov.uscourts.cand.407208.223.0_3.pdf","https://www.theverge.com/2024/8/13/24220231/stability-midjourney-deviantart-runway-andersen-copyright-class-action-allowed-proceed"]}
+{"id":"getty-v-stability-us","name":"Getty Images (US) v. Stability AI","plaintiff":"Getty Images (US), Inc.","defendants":["Stability AI, Inc.","Stability AI Ltd."],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California (refiled from D. Del.)","caseNumber":"3:25-cv-06891 (originally D. Del. 1:23-cv-00135)","filed":"2025-08-14","status":"active","stage":"discovery","claims":["copyright-infringement","trademark-infringement","unfair-competition","dmca-violation"],"summary":"Getty alleges Stable Diffusion was trained on millions of Getty images and reproduces the Getty watermark in some generations. Getty voluntarily dismissed its February 2023 Delaware case and refiled in N.D. Cal. on August 14, 2025; on April 23, 2026 Judge Trina Thompson denied most of Stability's motion to dismiss, dismissing only the DMCA false-CMI claim with leave to amend.","sources":["https://www.courtlistener.com/docket/66788385/getty-images-us-inc-v-stability-ai-inc/","https://ipde.com/blog/2025/08/15/getty-images-voluntarily-dismisses-ai-action-to-re-file-in-nd-cal/","https://chatgptiseatingtheworld.com/2026/04/24/getty-images-mostly-wins-surviving-motion-to-dismiss-claims-except-for-dmca-cmi-claim-but-can-amend-that-claim-in-light-of-judge-thompsons-ruling/"]}
+{"id":"getty-v-stability-uk","name":"Getty Images (UK) v. Stability AI Ltd.","plaintiff":"Getty Images (US), Inc.; Getty Images International (UK)","defendants":["Stability AI Ltd."],"jurisdiction":"UK (England and Wales)","court":"High Court of Justice, Business and Property Courts of England and Wales; Court of Appeal (pending)","caseNumber":"IL-2023-000007; [2025] EWHC 2863 (Ch)","filed":"2023-01","status":"judgment","stage":"appeal","claims":["copyright-infringement","trademark-infringement"],"summary":"After a June 2025 trial, Mrs Justice Joanna Smith handed down judgment on November 4, 2025: Getty had abandoned its training and output claims, its secondary copyright claim was dismissed because a model that does not store or reproduce the works is not an \"infringing copy\", and it won only limited, historic trade mark findings over watermarks in certain older model versions. In December 2025 the judge granted Getty permission to appeal the secondary infringement ruling and refused Stability permission to appeal the trade mark findings; no Court of Appeal hearing had taken place as of September 2026.","sources":["https://www.judiciary.uk/judgments/getty-images-v-stability-ai/","https://caselaw.nationalarchives.gov.uk/ewhc/ch/2025/3343","https://cms.law/en/gbr/legal-updates/getty-images-v-stability-ai-permission-to-appeal-sought-by-getty-images-at-consequentials-hearing"]}
+{"id":"disney-universal-v-midjourney","name":"Disney Enterprises & Universal City Studios v. Midjourney (consolidated with Warner Bros. v. Midjourney)","plaintiff":"Disney Enterprises, Marvel, Lucasfilm, Twentieth Century Fox, Universal City Studios, DreamWorks Animation; Warner Bros. Entertainment and affiliates (consolidated case)","defendants":["Midjourney, Inc."],"jurisdiction":"US (C.D. Cal.)","court":"United States District Court for the Central District of California","caseNumber":"2:25-cv-05275 (lead); 2:25-cv-08376 (consolidated)","filed":"2025-06-11","status":"active","stage":"discovery","claims":["copyright-infringement"],"summary":"First major Hollywood studio suit against a generative AI company, alleging Midjourney trained on and outputs copies of characters such as Darth Vader, Elsa, and the Minions. Midjourney answered in August 2025; Warner Bros. filed a similar suit in September 2025, and the court consolidated the cases for all purposes in November 2025. Training-data discovery disputes were before the magistrate judge in September 2026.","sources":["https://www.courtlistener.com/docket/70513159/disney-enterprises-inc-v-midjourney-inc/","https://www.reuters.com/business/media-telecom/disney-universal-sue-image-creator-midjourney-copyright-infringement-2025-06-11/"]}
+{"id":"studios-v-minimax","name":"Disney, Universal & Warner Bros. v. MiniMax (Hailuo AI)","plaintiff":"Disney Enterprises and affiliates; Universal City Studios and DreamWorks Animation; Warner Bros. Entertainment and affiliates","defendants":["MiniMax (Nanonoble Pte. Ltd.; Shanghai Xiyu Jizhi Technology Co. Ltd.)"],"jurisdiction":"US (C.D. Cal.)","court":"United States District Court for the Central District of California","caseNumber":"2:25-cv-08768","filed":"2025-09-16","status":"active","stage":"discovery","claims":["copyright-infringement"],"summary":"Three studios allege China-based MiniMax built its Hailuo image and video generator on their copyrighted characters and markets it as a \"Hollywood studio in your pocket\". The court denied the defendants' motions to dismiss for lack of personal jurisdiction and failure to state a claim on May 22, 2026; the defendants answered with a counterclaim in June 2026.","sources":["https://www.reuters.com/legal/litigation/disney-universal-warner-bros-discovery-sue-chinas-minimax-copyright-infringement-2025-09-16/","https://variety.com/2025/digital/news/disney-warner-bros-discovery-nbcu-lawsuit-minimax-chinese-ai-company-1236520395/"]}
+{"id":"kneschke-v-laion","name":"Kneschke v. LAION e.V.","plaintiff":"Robert Kneschke (photographer)","defendants":["LAION e.V."],"jurisdiction":"Germany (EU)","court":"Bundesgerichtshof (Federal Court of Justice), I. Zivilsenat; previously Landgericht Hamburg and Hanseatisches Oberlandesgericht Hamburg","caseNumber":"I ZR 281/25 (BGH); 5 U 104/24 (OLG Hamburg); 310 O 227/23 (LG Hamburg)","filed":"2023","status":"active","stage":"appeal","claims":["copyright-infringement"],"summary":"A photographer sued LAION over the download of his photo while building the LAION-5B image-text dataset. The Hamburg Regional Court dismissed the claim in September 2024 under the scientific research text and data mining exception, and on December 10, 2025 the Hamburg Higher Regional Court dismissed his appeal, holding that both the research and general TDM exceptions applied and that his natural-language opt-out was not shown to be machine-readable in 2021. The Federal Court of Justice heard the further appeal on September 3, 2026 and scheduled its decision for December 17, 2026.","sources":["https://www.bundesgerichtshof.de/SharedDocs/Termine/DE/Termine/IZR281-25.html","https://www.twobirds.com/en/insights/2025/germany/higher-regional-court-hamburg-confirms-ai-training-was-permitted-(kneschke-v,-d-,-laion)"]}
+{"id":"concord-v-anthropic","name":"Concord Music Group v. Anthropic","plaintiff":"Concord Music Group, Universal Music Publishing Group, ABKCO, and other music publishers","defendants":["Anthropic PBC"],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California (transferred from M.D. Tenn.)","caseNumber":"5:24-cv-03811 (originally M.D. Tenn. 3:23-cv-01092)","filed":"2023-10-18","status":"active","stage":"summary-judgment","claims":["copyright-infringement","dmca-violation"],"summary":"Music publishers allege Claude was trained on copyrighted song lyrics and outputs them on request. The case moved to N.D. Cal. in June 2024; the parties stipulated to output guardrails in January 2025, and Judge Eumi Lee denied a preliminary injunction on March 25, 2025. Cross-motions for summary judgment were pending as of September 2026, when later suits by Round Hill Music and by Sony Music Publishing and Warner Chappell were related to this case.","sources":["https://storage.courtlistener.com/recap/gov.uscourts.cand.431519/gov.uscourts.cand.431519.321.0.pdf","https://www.billboard.com/business/legal/anthropic-claude-ai-music-publishers-lyrics-lawsuit-1235427547/"]}
+{"id":"concord-v-anthropic-ii","name":"Concord Music Group v. Anthropic, Amodei & Mann (piracy claims)","plaintiff":"Music publishers led by Concord Music Group","defendants":["Anthropic PBC","Dario Amodei","Benjamin Mann"],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California","caseNumber":"5:26-cv-00880","filed":"2026-01-28","status":"active","stage":"motion-to-dismiss","claims":["copyright-infringement"],"summary":"After the court denied leave to add piracy allegations to the first Concord case, the publishers filed a second suit alleging Anthropic obtained song lyrics through piracy, covering more than 20,000 songs by their count. Anthropic and Amodei filed partial motions to dismiss in August 2026, with a hearing set for November 4, 2026.","sources":["https://www.reuters.com/legal/litigation/anthropic-faces-new-music-publisher-lawsuit-over-alleged-piracy-2026-01-28/"]}
+{"id":"sony-warner-chappell-v-anthropic","name":"Sony Music Publishing & Warner Chappell Music v. Anthropic, Amodei & Mann","plaintiff":"Sony Music Publishing (US) LLC; Warner Chappell Music, Inc.; affiliated publishers","defendants":["Anthropic PBC","Dario Amodei","Benjamin Mann"],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California","caseNumber":"5:26-cv-09217","filed":"2026-08-28","status":"active","stage":"complaint","claims":["copyright-infringement","dmca-violation"],"summary":"The publishing arms of Sony Music and Warner Music allege direct infringement by torrenting against Anthropic and two co-founders, contributory infringement against the co-founders, and direct infringement and removal of copyright management information against Anthropic. On September 8, 2026 the case was related to Concord Music Group v. Anthropic and reassigned to Judge Eumi Lee.","sources":["https://www.musicbusinessworldwide.com/now-sony-music-publishing-and-warner-chappell-sue-anthropic-in-multi-billion-dollar-lawsuit-one-of-the-largest-and-most-blatant-ongoing-thefts-of-intellectual-property-in-history/","https://www.reuters.com/legal/government/sony-warner-music-sue-anthropic-over-songs-used-ai-training-2026-08-31/"]}
+{"id":"riaa-v-suno","name":"UMG Recordings et al. v. Suno (RIAA-coordinated)","plaintiff":"UMG Recordings, Capitol Records, Sony Music Entertainment and affiliated labels (Warner labels dismissed after settling)","defendants":["Suno, Inc."],"jurisdiction":"US (D. Mass.)","court":"United States District Court for the District of Massachusetts","caseNumber":"1:24-cv-11611","filed":"2024-06-24","status":"active","stage":"discovery","claims":["copyright-infringement"],"summary":"Major labels allege Suno trained its music-generation model on copyrighted recordings. Warner Music Group settled with Suno in November 2025 and its labels were dismissed; UMG and Sony remain in the case, with fact discovery scheduled to close September 30, 2026. Companion case against Udio proceeds in S.D.N.Y.","sources":["https://www.riaa.com/wp-content/uploads/2024/06/UMG-Recordings-et-al-v-Suno-Inc-et-al-Complaint.pdf","https://www.reuters.com/legal/litigation/warner-music-group-settles-copyright-case-with-suno-licensed-ai-music-2025-11-25/"]}
+{"id":"riaa-v-udio","name":"UMG Recordings et al. v. Uncharted Labs (Udio)","plaintiff":"Sony Music Entertainment and Arista Records (UMG, Capitol, and Warner labels dismissed after settling)","defendants":["Uncharted Labs, Inc. (d/b/a Udio)"],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:24-cv-04777","filed":"2024-06-24","status":"active","stage":"discovery","claims":["copyright-infringement"],"summary":"Companion case to the Suno suit with the same training-data theory. UMG settled with Udio in October 2025 and Warner Music Group in November 2025, and their claims were dismissed; Sony's claims continue, and on June 29, 2026 the court denied the remaining plaintiffs' motion to amend the complaint.","sources":["https://www.riaa.com/wp-content/uploads/2024/06/Sony-Music-et-al-v-Uncharted-Labs-dba-Udio-Complaint.pdf","https://apnews.com/article/udio-suno-ai-music-universal-b90f9f5f968101ef617e41c5369da02a","https://investors.wmg.com/news-events/news-releases/news-details/2025/WARNER-MUSIC-GROUP-AND-UDIO-COLLABORATE-TO-BUILD-A-NEW-LICENSED-MUSIC-CREATION-SERVICE/default.aspx"]}
+{"id":"gema-v-openai","name":"GEMA v. OpenAI","plaintiff":"GEMA (German music rights collecting society)","defendants":["Two OpenAI group companies"],"jurisdiction":"Germany (EU)","court":"Landgericht Muenchen I (Munich Regional Court I), 42nd Civil Chamber; appeal to Oberlandesgericht Muenchen","caseNumber":"42 O 14139/24 (appeal 6 U 3662/25 e)","filed":"2024","status":"judgment","stage":"appeal","claims":["copyright-infringement"],"summary":"On November 11, 2025 the Munich Regional Court largely granted GEMA's claims for an injunction, information, and damages over the lyrics of nine German songs, while dismissing claims based on personality rights. According to the published headnotes, memorization of lyrics in the model is a reproduction, the text and data mining exception covers compiling training data but not the training itself, and the operators are liable for infringing outputs. The judgment is not final; OpenAI has appealed to the Munich Higher Regional Court.","sources":["https://www.justiz.bayern.de/gerichte-und-behoerden/landgericht/muenchen-1/presse/2025/11.php","https://www.gesetze-bayern.de/Content/Document/Y-300-Z-GRURRS-B-2025-N-30204"]}
+{"id":"gema-v-suno","name":"GEMA v. Suno","plaintiff":"GEMA (German music rights collecting society)","defendants":["Suno, Inc."],"jurisdiction":"Germany (EU)","court":"Landgericht Muenchen I (Munich Regional Court I), 42nd Civil Chamber","caseNumber":"42 O 763/25","filed":"2025-01-21","status":"judgment","stage":"closed","claims":["copyright-infringement"],"summary":"On July 31, 2026 the Munich Regional Court mostly granted GEMA's claims for an injunction, information, and damages over six well-known songs, including \"Daddy Cool\" and \"Forever Young\", finding infringing reproductions in Suno's US training (where it held fair use did not apply), in models stored in Germany, and in outputs. The court's press release says Suno used stream-ripping to extract the works from YouTube, circumventing a technical protection measure. The first-instance judgment is not final, and Suno said it was reviewing its options, including an appeal.","sources":["https://www.justiz.bayern.de/gerichte-und-behoerden/landgericht/muenchen-1/presse/2026/16.php","https://www.tagesschau.de/kultur/urteil-gema-vs-suno-100.html"]}
+{"id":"lehrman-v-lovo","name":"Lehrman v. Lovo","plaintiff":"Paul Skye Lehrman, Linnea Sage (voice actors)","defendants":["Lovo, Inc."],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:24-cv-03770","filed":"2024-05-16","status":"active","stage":"discovery","claims":["right-of-publicity","consumer-protection","breach-of-contract","copyright-infringement"],"summary":"Voice actors allege their voices were cloned and sold as text-to-speech products without consent. On July 10, 2025 Judge Oetken let the New York Civil Rights Law, consumer protection, contract, and one copyright claim over an original recording proceed, dismissed the Lanham Act and several other claims, and allowed repleading of the training-based copyright claim. The case was stayed on May 28, 2026 after Lovo filed a suggestion of bankruptcy, before a ruling on the renewed motion to dismiss.","sources":["https://www.courtlistener.com/docket/68535780/lehrman-v-lovo-inc/","https://nysd.uscourts.gov/sites/default/files/2025-07/Lovo%20v%20Lehrman.pdf"]}
+{"id":"alvarez-v-meta","name":"Alvarez et al. v. Meta Platforms (face recognition and AI training)","plaintiff":"Francisco Alvarez, Jeremy Wahl, and their minor children (putative nationwide, Illinois, and California classes)","defendants":["Meta Platforms, Inc."],"jurisdiction":"US (N.D. Ill.)","court":"United States District Court for the Northern District of Illinois","caseNumber":"1:26-cv-10773","filed":"2026-09-04","status":"active","stage":"complaint","claims":["privacy","right-of-publicity"],"summary":"Parents and children in Illinois and California allege Meta used Facebook and Instagram photos without consent to build face recognition for its smart glasses and to train generative AI models, in violation of the Illinois Biometric Information Privacy Act and California privacy and right of publicity law. Meta called the suit without merit.","sources":["https://news.bloomberglaw.com/litigation/meta-sued-over-smart-glasses-collection-of-user-biometric-data","https://www.wired.com/story/meta-sued-over-training-data-for-its-ai-and-face-recognition-systems/"]}
+{"id":"reddit-v-anthropic","name":"Reddit, Inc. v. Anthropic PBC","plaintiff":"Reddit, Inc.","defendants":["Anthropic PBC"],"jurisdiction":"US (California state court)","court":"Superior Court of California, County of San Francisco","caseNumber":"CGC-25-625892","filed":"2025-06-04","status":"active","stage":"complaint","claims":["breach-of-contract","unjust-enrichment","unfair-competition","tort"],"summary":"Reddit alleges Anthropic used Reddit content without permission and in breach of Reddit's user agreement. Anthropic removed the case to federal court, but in March 2026 Judge Trina Thompson remanded it to San Francisco Superior Court, where it proceeds.","sources":["https://www.courthousenews.com/wp-content/uploads/2026/03/reddit-v-anthropic-remand-order.pdf","https://www.courthousenews.com/reddit-privacy-case-against-anthropic-kicked-back-to-state-court/"]}
+{"id":"reddit-v-serpapi-perplexity","name":"Reddit, Inc. v. SerpApi, Oxylabs, AWMProxy & Perplexity AI","plaintiff":"Reddit, Inc.","defendants":["SerpApi LLC","Oxylabs UAB","AWMProxy","Perplexity AI, Inc."],"jurisdiction":"US (S.D.N.Y.)","court":"United States District Court for the Southern District of New York","caseNumber":"1:25-cv-08736","filed":"2025-10-22","status":"active","stage":"discovery","claims":["dmca-violation","tort"],"summary":"Reddit brings DMCA anti-circumvention and related claims against scraping services and Perplexity over the harvesting of Reddit content. On July 31, 2026 Judge Engelmayer largely denied the motions to dismiss, letting the DMCA anti-circumvention and civil conspiracy claims proceed while dismissing the unfair competition and unjust enrichment claims and one DMCA claim against SerpApi.","sources":["https://www.reuters.com/legal/litigation/perplexity-ai-loses-bid-toss-reddit-lawsuit-over-data-scraping-2026-07-31/"]}
+{"id":"amazon-v-perplexity","name":"Amazon.com Services v. Perplexity AI (Comet shopping agent)","plaintiff":"Amazon.com Services, LLC","defendants":["Perplexity AI, Inc."],"jurisdiction":"US (N.D. Cal.; 9th Cir.)","court":"United States District Court for the Northern District of California","caseNumber":"3:25-cv-09514 (9th Cir. No. 26-1444)","filed":"2025-11","status":"active","stage":"discovery","claims":["tort"],"summary":"Amazon alleges Perplexity's Comet browser agent accessed password-protected parts of Amazon in violation of federal and California computer-access laws. A preliminary injunction granted on March 9, 2026 was vacated by a published Ninth Circuit opinion on August 4, 2026, which found Amazon unlikely to show that Perplexity \"accessed\" its computers under the CFAA, and the case was remanded.","sources":["https://cdn.ca9.uscourts.gov/datastore/opinions/2026/08/04/26-1444.pdf","https://www.geekwire.com/2026/judge-blocks-perplexitys-ai-bot-from-shopping-on-amazon-in-early-test-of-agentic-commerce/"]}
+{"id":"musk-v-altman","name":"Musk v. Altman, et al.","plaintiff":"Elon Musk","defendants":["Samuel Altman","OpenAI, Inc. and affiliated entities"],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California","caseNumber":"4:24-cv-04722","filed":"2024-08-05","status":"active","stage":"trial","claims":["breach-of-contract","unjust-enrichment"],"summary":"Musk sued OpenAI and Altman over OpenAI's conduct and structure, with claims including breach of charitable trust. After a spring 2026 trial, an advisory jury found on May 18, 2026 that the tried claims were barred by the statute of limitations, and Judge Gonzalez Rogers adopted that result in a May 20 post-trial order. Remaining claims and OpenAI's counterclaims are still being briefed, a mediator was appointed in July 2026, and no final judgment had been entered as of August 2026.","sources":["https://www.docketbird.com/court-documents/Musk-v-Altman-et-al/POST-TRIAL-ORDER-RE-ADVISORY-VERDICT-Signed-by-Judge-Yvonne-Gonzalez-Rogers-on-5-20-2026/cand-4:2024-cv-04722-00580","https://www.cnbc.com/2026/05/18/musk-altman-openai-trial-verdict.html","https://www.npr.org/2026/05/18/nx-s1-5822366/musk-altman-openai-jury-verdict-claims-dismissed"]}
+{"id":"xai-v-openai","name":"X.AI Corp. v. OpenAI (trade secrets)","plaintiff":"X.AI Corp.","defendants":["OpenAI, Inc. and affiliated entities"],"jurisdiction":"US (N.D. Cal.; 9th Cir.)","court":"United States District Court for the Northern District of California","caseNumber":"3:25-cv-08133 (9th Cir. No. 26-4486)","filed":"2025-09-24","status":"dismissed","stage":"appeal","claims":["unfair-competition"],"summary":"xAI alleged OpenAI misappropriated its trade secrets. Judge Rita Lin dismissed the amended complaint with leave to amend in February 2026, then dismissed the case without leave to amend on June 15, 2026; xAI appealed to the Ninth Circuit in July 2026.","sources":["https://www.courthousenews.com/judge-tosses-xai-claims-that-openai-stole-trade-secrets/","https://beckreedriden.com/xai-v-openai-motion-to-dismiss-granted-but-the-story-might-not-be-over/"]}
+{"id":"x-corp-v-apple-openai","name":"X Corp. & xAI v. Apple and OpenAI (antitrust)","plaintiff":"X Corp.; X.AI LLC (recaptioned SpaceXAI LLC in August 2026)","defendants":["Apple Inc.","OpenAI, Inc. and affiliated entities"],"jurisdiction":"US (N.D. Tex.)","court":"United States District Court for the Northern District of Texas (Fort Worth)","caseNumber":"4:25-cv-00914","filed":"2025-08-25","status":"active","stage":"discovery","claims":["antitrust","unfair-competition"],"summary":"X and xAI bring antitrust, unfair competition, and civil conspiracy claims against Apple and OpenAI over their partnership. Judge Pittman denied the motions to dismiss on November 13, 2025, and an April 2026 amended scheduling order moved trial to January 11, 2027.","sources":["https://www.courtlistener.com/docket/71191818/x-corp-v-apple-inc/","https://cdn.arstechnica.net/wp-content/uploads/2026/05/X-Corp-v-Apple-OpenAI-Order-5-13-26.pdf"]}
+{"id":"anthropic-v-department-of-war","name":"Anthropic PBC v. U.S. Department of War (supply-chain-risk designation)","plaintiff":"Anthropic PBC","defendants":["U.S. Department of War","other federal officials and agencies"],"jurisdiction":"US (N.D. Cal.; D.C. Cir.)","court":"United States District Court for the Northern District of California; U.S. Court of Appeals for the D.C. Circuit (separate petition)","caseNumber":"3:26-cv-01996 (N.D. Cal.); D.C. Cir. No. 26-1049","filed":"2026-03-09","status":"judgment","stage":"closed","claims":["regulatory-investigation"],"summary":"Anthropic challenged the Pentagon's designation of the company as a supply-chain risk. Judge Rita Lin issued a preliminary injunction on March 26, 2026, decided cross-motions for summary judgment and entered judgment on August 27, 2026, and press coverage reported the ruling found the 10 U.S.C. 3252 designation unlawful. A separate D.C. Circuit petition over a related designation was argued May 19, 2026 and remained undecided as of late August 2026.","sources":["https://storage.courtlistener.com/recap/gov.uscourts.cand.465515/gov.uscourts.cand.465515.252.0.pdf","https://www.npr.org/2026/03/26/nx-s1-5762971/judge-temporarily-blocks-anthropic-ban","https://www.nextgov.com/artificial-intelligence/2026/08/judge-rules-anthropic-supply-chain-risk-designation-was-illegal-and-baseless/415698/"]}
+{"id":"florida-ag-v-openai","name":"State of Florida (Office of the Attorney General) v. OpenAI & Sam Altman","plaintiff":"State of Florida, Office of the Attorney General","defendants":["OpenAI entities","Samuel Altman"],"jurisdiction":"US (Florida state court)","court":"Circuit Court of the Tenth Judicial Circuit, Highlands County, Florida","caseNumber":"28-2026-CA-000295","filed":"2026-06-01","status":"active","stage":"complaint","claims":["consumer-protection","tort"],"summary":"Florida's attorney general brought a 10-count suit alleging OpenAI and Altman violated the Florida Deceptive and Unfair Trade Practices Act, with negligence, gross negligence, strict liability, and public nuisance claims. OpenAI removed the case to federal court in July 2026, and it was remanded to state court on September 8, 2026.","sources":["https://www.myfloridalegal.com/sites/default/files/openai-filed-stamped-complaint.pdf","https://www.abajournal.com/news/article/florida-ag-brings-first-in-the-nation-state-led-lawsuit-against-openai-and-its-ceo","https://www.law.com/dailybusinessreview/2026/09/09/judge-gives-florida-a-win-in-sending-openai-lawsuit-back-to-state-court/"]}
+{"id":"ftc-ai-companion-chatbot-inquiry","name":"FTC 6(b) inquiry: AI chatbots acting as companions","plaintiff":"US Federal Trade Commission (Section 6(b) study)","defendants":["Alphabet","Character Technologies","Instagram","Meta","OpenAI","Snap","xAI"],"jurisdiction":"US (federal regulatory)","court":"Federal Trade Commission","caseNumber":null,"filed":"2025-09-11","status":"active","stage":"discovery","claims":["regulatory-investigation"],"summary":"The FTC issued Section 6(b) orders to seven companies seeking information about consumer AI chatbots that act as companions. Section 6(b) studies gather information without a specific law enforcement purpose; no staff report had been published as of September 2026. The FTC's earlier 6(b) study of AI partnerships and investments (Microsoft, Alphabet, Amazon, OpenAI, Anthropic) concluded with a staff report in January 2025.","sources":["https://www.ftc.gov/news-events/news/press-releases/2025/09/ftc-launches-inquiry-ai-chatbots-acting-companions","https://www.ftc.gov/news-events/news/press-releases/2025/01/ftc-issues-staff-report-ai-partnerships-investments-study"]}
+{"id":"garcia-v-character-ai","name":"Garcia v. Character Technologies","plaintiff":"Megan Garcia (estate of Sewell Setzer III)","defendants":["Character Technologies, Inc.","Noam Shazeer","Daniel De Freitas","Google LLC"],"jurisdiction":"US (M.D. Fla.)","court":"United States District Court for the Middle District of Florida","caseNumber":"6:24-cv-01903","filed":"2024-10-22","status":"settled","stage":"closed","claims":["tort"],"summary":"Wrongful-death and product-liability suit alleging Character.AI's chatbot contributed to a 14-year-old's suicide. In May 2025 Judge Conway granted the motions to dismiss in part and denied them in part, declining at that stage to hold that the chatbot's output was protected speech. The parties notified the court of a settlement and the case was dismissed on January 7, 2026; reports said Character.AI and Google were finalizing settlements in related family suits, and terms were not disclosed.","sources":["https://molawyersmedia.com/2026/01/09/google-character-ai-chatbot-lawsuit-settlement/","https://www.leagle.com/decision/infdco20260612e78","https://caselaw.findlaw.com/court/us-dis-crt-m-d-flo-orl-div/117299600.html"]}
+{"id":"chatgpt-product-liability-jccp","name":"ChatGPT Product Liability Cases (JCCP No. 5431), including Raine v. OpenAI","plaintiff":"Matthew and Maria Raine and other individual plaintiffs and estates","defendants":["OpenAI, Inc. and affiliated entities","Microsoft Corporation (in some coordinated cases)"],"jurisdiction":"US (California state court)","court":"Superior Court of California, County of San Francisco (coordination proceeding)","caseNumber":"JCCP No. 5431 (Raine: CGC-25-628528)","filed":"2025-08-26","status":"consolidated","stage":"complaint","claims":["tort"],"summary":"Wrongful-death and product-liability suits against OpenAI over ChatGPT, beginning with Raine v. OpenAI in August 2025 and a wave of seven suits filed in November 2025. On February 3, 2026 the cases were coordinated in San Francisco as the ChatGPT Product Liability Cases, and in August 2026 Judge Ethan Schulman approved a plaintiffs' leadership structure.","sources":["https://reason.com/wp-content/uploads/2026/06/chatgpt-product-liability-cases-coordination.pdf","https://www.mlex.com/mlex/articles/2510216/san-francisco-judge-approves-leadership-team-for-chatgpt-product-liability-cases"]}
+{"id":"tumbler-ridge-v-openai","name":"Tumbler Ridge shooting suits against OpenAI & Sam Altman","plaintiff":"Victims and families connected to the Tumbler Ridge mass shooting","defendants":["OpenAI entities","Samuel Altman"],"jurisdiction":"US (N.D. Cal.)","court":"United States District Court for the Northern District of California","caseNumber":"Multiple, e.g. 3:26-cv-03702 (Schofield v. Altman), 3:26-cv-09377 (Hoekstra v. Altman)","filed":"2026-04-29","status":"active","stage":"complaint","claims":["tort"],"summary":"Seven suits filed in April 2026 and 30 more in September 2026 accuse OpenAI and CEO Sam Altman of negligence, aiding and abetting a mass shooting, and product liability.","sources":["https://www.theguardian.com/world/2026/sep/02/openai-lawsuits-tumbler-ridge-mass-shooting","https://techcrunch.com/2026/09/02/openai-faces-30-more-lawsuits-tied-to-tumbler-ridge-shooting/"]}

2026-10-12/ai-policy.jsonl ADDED Viewed

	@@ -0,0 +1,10 @@

+{"id":"eu-ai-act","name":"EU AI Act","jurisdiction":"European Union","status":"active","type":"regulation","enactedDate":"2024-08-01","milestones":[{"date":"2024-08-01","event":"Entered into force"},{"date":"2025-02-02","event":"Prohibited practices banned (social scoring, manipulation, untargeted scraping)"},{"date":"2025-08-02","event":"GPAI / foundation-model obligations active (transparency, copyright, model cards)"},{"date":"2026-07-27","event":"Digital Omnibus on AI (Regulation (EU) 2026/1744) in force, deferring high-risk obligations"},{"date":"2026-08-02","event":"Article 50 transparency duties and the general body of the Act apply"},{"date":"2027-12-02","event":"Stand-alone high-risk systems (Annex III) compliance deadline, as deferred by the Digital Omnibus"},{"date":"2028-08-02","event":"High-risk AI embedded in regulated products (Annex I) compliance deadline, as deferred by the Digital Omnibus"}],"scope":"Any AI system placed on the EU market or affecting EU users. Tiered by risk level: prohibited, high-risk, limited, minimal.","summary":"Tiered risk-based AI regulation. GPAI providers must publish summaries of training data, respect copyright opt-outs, evaluate systemic risk for very capable models. High-risk systems require conformity assessments + EU database registration.","lead":"European Commission AI Office + national authorities","penalties":"Up to €35M or 7% of global turnover for prohibited use; €15M or 3% for non-compliance with GPAI obligations","url":"https://artificialintelligenceact.eu"}
+{"id":"us-ai-eo-trump-2025","name":"Removing Barriers to American Leadership in AI (EO 14179)","jurisdiction":"United States (federal)","status":"active","type":"executive-order","enactedDate":"2025-01-23","milestones":[{"date":"2025-01-20","event":"Biden EO 14110 revoked by EO 14148"},{"date":"2025-01-23","event":"Signed by President Trump; orders a review of actions taken under the revoked EO 14110"},{"date":"2025-07-23","event":"AI Action Plan released by White House OSTP"}],"scope":"All US federal agencies; sets priorities for AI competitiveness, safety, and procurement.","summary":"Issued after Biden's 2023 EO 14110 was revoked by EO 14148. Focuses on accelerating US AI competitiveness and reducing perceived regulatory barriers. NIST's AI Safety Institute was reorganized as the Center for AI Standards and Innovation (CAISI) in June 2025. State-level laws (CA, CO, NY, IL) increasingly fill the federal gap.","lead":"White House OSTP + NIST CAISI","penalties":"No direct penalties (executive order applies to agencies, not private actors)","url":"https://www.whitehouse.gov/presidential-actions/2025/01/removing-barriers-to-american-leadership-in-artificial-intelligence/"}
+{"id":"guard-act-2026","name":"GUARD Act","jurisdiction":"United States (federal)","status":"proposed","type":"law","enactedDate":null,"milestones":[{"date":"2025-10-28","event":"Introduced as S.3062 by Sens. Hawley (R) and Blumenthal (D)"},{"date":"2026-04-30","event":"Senate Judiciary Committee advanced 22-0 (unanimous)"},{"date":"2026-05-11","event":"Placed on the Senate legislative calendar; no floor vote yet"}],"scope":"AI chatbots interacting with US users. Special focus on minors.","summary":"Government-ID age verification for AI chatbots, flat ban on AI companions for minors, mandatory non-human disclosures every 30 minutes, criminal penalties for design choices that knowingly route minors into sexually explicit or self-harm content.","lead":"Senate Judiciary Committee; FTC implementation expected","penalties":"Civil + criminal penalties for design choices; product-recall authority","url":"https://www.judiciary.senate.gov"}
+{"id":"ca-sb-1047","name":"California SB 1047","jurisdiction":"California","status":"vetoed","type":"law","enactedDate":null,"milestones":[{"date":"2024-09-29","event":"Vetoed by Governor Newsom"},{"date":"2025-09-29","event":"Successor bill SB 53 (Transparency in Frontier Artificial Intelligence Act) signed"}],"scope":"AI models trained with > $100M compute or 10^26 FLOPs. Frontier AI labs operating in California.","summary":"Would have required pre-deployment safety testing, kill-switch capability, and whistleblower protections for frontier AI. Vetoed in 2024. Its successor, SB 53, signed 2025-09-29, requires large frontier developers to publish safety frameworks and report critical safety incidents instead of mandating pre-deployment testing.","lead":"California Office of Technology + AG (would have been)","penalties":"N/A (vetoed)","url":"https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240SB1047"}
+{"id":"ca-ab-2013","name":"California AB 2013 (Generative AI Training Data Transparency)","jurisdiction":"California","status":"active","type":"law","enactedDate":"2024-09-28","milestones":[{"date":"2024-09-28","event":"Signed"},{"date":"2026-01-01","event":"Compliance deadline"}],"scope":"Developers of generative AI systems made available to California residents.","summary":"Requires public disclosure of training-data summaries, data sources, copyright considerations, and PII content for any generative AI system available to Californians starting January 2026.","lead":"California AG","penalties":"Civil enforcement; specific penalties TBD","url":"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240AB2013"}
+{"id":"uk-aisi","name":"UK AI Security Institute","jurisdiction":"United Kingdom","status":"active","type":"voluntary-framework","enactedDate":"2023-11","milestones":[{"date":"2023-11-02","event":"Founded at AI Safety Summit"},{"date":"2024-05-21","event":"Voluntary pre-deployment testing agreements with frontier labs"},{"date":"2025-02-14","event":"Renamed from AI Safety Institute to AI Security Institute"}],"scope":"Frontier AI labs operating in UK or releasing to UK users. Voluntary at present; may become mandatory.","summary":"Pre-deployment red-team evaluations of frontier models across cyber, biosecurity, autonomous-system, political-influence dimensions. Operates under DSIT (Department for Science, Innovation and Technology).","lead":"DSIT / UK AI Security Institute","penalties":"No statutory penalties; testing access is voluntary","url":"https://www.aisi.gov.uk"}
+{"id":"china-genai-measures","name":"Interim Measures for the Management of Generative AI Services","jurisdiction":"China","status":"active","type":"regulation","enactedDate":"2023-08-15","milestones":[{"date":"2023-08-15","event":"Entered into force"},{"date":"2024-04-01","event":"Mandatory generative AI service registration with CAC"},{"date":"2025-09-01","event":"AI-generated content labeling rules effective"}],"scope":"Generative AI services offered to the Chinese public. Registration, licensing, and content controls.","summary":"Pre-launch security assessment, content filtering aligned with Chinese law, watermarking + labeling of AI-generated content, training-data provenance documentation.","lead":"Cyberspace Administration of China (CAC)","penalties":"License revocation, fines, criminal referral for violations","url":"https://www.cac.gov.cn"}
+{"id":"korea-basic-ai-act","name":"Korea Basic Act on AI","jurisdiction":"South Korea","status":"active","type":"law","enactedDate":"2024-12-26","milestones":[{"date":"2024-12-26","event":"Passed National Assembly"},{"date":"2026-01-22","event":"Entered into force, with a grace period of at least one year before fines are imposed"}],"scope":"AI providers operating in South Korea or providing services to Korean users.","summary":"Comprehensive AI governance framework. Risk-based tiers (similar to EU AI Act lite). Korean Ministry of Science and ICT designated as enforcement authority. Lighter-touch than EU AI Act on GPAI obligations.","lead":"Ministry of Science and ICT (MSIT)","penalties":"Administrative fines up to KRW 30M ($22k) per violation","url":"https://www.msit.go.kr"}
+{"id":"nist-ai-rmf","name":"NIST AI Risk Management Framework","jurisdiction":"United States (voluntary, internationally referenced)","status":"active","type":"voluntary-framework","enactedDate":"2023-01-26","milestones":[{"date":"2023-01-26","event":"AI RMF 1.0 published"},{"date":"2024-07-26","event":"Generative AI Profile published"}],"scope":"Voluntary; widely referenced in US federal procurement and state laws.","summary":"Risk-management lifecycle for AI systems: Govern, Map, Measure, Manage. Generative AI profile addresses model-specific risks. Foundation for many state and sector-specific AI laws.","lead":"NIST","penalties":"Voluntary (no penalties)","url":"https://www.nist.gov/itl/ai-risk-management-framework"}
+{"id":"iso-42001","name":"ISO/IEC 42001 (AI Management Systems)","jurisdiction":"International (ISO standard)","status":"active","type":"standard","enactedDate":"2023-12-18","milestones":[{"date":"2023-12-18","event":"Published"},{"date":"2024-Q4","event":"First commercial certifications issued"}],"scope":"Organizations developing or deploying AI systems. Voluntary certification.","summary":"First international management-system standard for AI. Specifies requirements for establishing, implementing, maintaining, and improving an AI management system. Often required by enterprise buyers.","lead":"ISO/IEC JTC 1/SC 42","penalties":"Voluntary; loss of certification on non-compliance","url":"https://www.iso.org/standard/81230.html"}

2026-10-12/attention.jsonl ADDED Viewed

	@@ -0,0 +1,12 @@

+{"id":"nvidia","name":"NVIDIA","news_24h":3,"news_7d":7,"trending_repos":0,"agent_hits":0,"raw_score":19,"attention_score":100,"rank":1,"top_articles":[{"title":"Show HN: Reflex – A GGUF-native Rust/CUDA engine for scale-to-zero LLM routing","source":"Hacker News AI","published_at":"2026-10-12T16:05:06.000Z"},{"title":"OCP Global Summit: Amid Ecosystem Adoption, NVIDIA Opens Vera Rubin Architecture to Help Scale AI Factories Worldwide","source":"NVIDIA AI Blog","published_at":"2026-10-12T15:00:25.000Z"},{"title":"Microsoft’s new AI chip makes it less reliant on Nvidia’s GPUs","source":"The Verge AI","published_at":"2026-10-12T13:00:00.000Z"}]}
+{"id":"anthropic","name":"Anthropic","news_24h":1,"news_7d":6,"trending_repos":4,"agent_hits":0,"raw_score":18,"attention_score":94.7,"rank":2,"top_articles":[{"title":"I'm an AI Founder (Claude Opus 5.5 music video)","source":"Hacker News AI","published_at":"2026-10-12T16:10:34.000Z"},{"title":"Anthropic is cutting off its internal evaluations from the internet","source":"The Verge AI","published_at":"2026-10-10T14:41:16.000Z"},{"title":"Anthropic can’t reliably control its AI agents — it’s cutting off its internal evals from the live internet instead","source":"TechCrunch AI","published_at":"2026-10-10T00:18:32.000Z"}]}
+{"id":"openai","name":"OpenAI","news_24h":0,"news_7d":5,"trending_repos":2,"agent_hits":0,"raw_score":9,"attention_score":47.4,"rank":3,"top_articles":[{"title":"These execs think voice AI hasn’t reached its ChatGPT moment yet","source":"TechCrunch AI","published_at":"2026-10-11T14:00:00.000Z"},{"title":"AI agent makers are promising privacy — will they deliver?","source":"The Verge AI","published_at":"2026-10-10T13:00:00.000Z"},{"title":"‘Pure insanity’: Mathematicians will need years to make sense of OpenAI’s latest drop","source":"The Verge AI","published_at":"2026-10-09T19:09:44.000Z"}]}
+{"id":"xai","name":"xAI","news_24h":1,"news_7d":2,"trending_repos":0,"agent_hits":0,"raw_score":6,"attention_score":31.6,"rank":4,"top_articles":[{"title":"Faithfulness Without Anchors: A Position Paper on the Structural Misalignments of Explainable AI Faithfulness Metrics for Feature Importance","source":"arXiv cs.AI","published_at":"2026-10-12T04:00:00.000Z"},{"title":"These Researchers Made AI Drive a Toyota Corolla to Get In-N-Out","source":"WIRED AI","published_at":"2026-10-07T18:45:00.000Z"}]}
+{"id":"huggingface","name":"Hugging Face","news_24h":0,"news_7d":5,"trending_repos":0,"agent_hits":0,"raw_score":5,"attention_score":26.3,"rank":5,"top_articles":[{"title":"Impactful scheduling for GPU clusters","source":"Hugging Face Blog","published_at":"2026-10-09T15:20:29.000Z"},{"title":"The model that didn't exist, so you made it yourself","source":"Hugging Face Blog","published_at":"2026-10-08T00:00:00.000Z"},{"title":"Multimodal open d1 decision models for the edge","source":"Hugging Face Blog","published_at":"2026-10-07T16:54:33.000Z"}]}
+{"id":"google","name":"Google","news_24h":0,"news_7d":1,"trending_repos":1,"agent_hits":0,"raw_score":3,"attention_score":15.8,"rank":6,"top_articles":[{"title":"Introducing Playground: Create and play custom games","source":"Google AI Blog","published_at":"2026-10-07T12:00:00.000Z"}]}
+{"id":"deepseek","name":"DeepSeek","news_24h":0,"news_7d":0,"trending_repos":1,"agent_hits":0,"raw_score":2,"attention_score":10.5,"rank":7,"top_articles":[]}
+{"id":"meta","name":"Meta","news_24h":0,"news_7d":0,"trending_repos":0,"agent_hits":0,"raw_score":0,"attention_score":0,"rank":8,"top_articles":[]}
+{"id":"mistral","name":"Mistral","news_24h":0,"news_7d":0,"trending_repos":0,"agent_hits":0,"raw_score":0,"attention_score":0,"rank":9,"top_articles":[]}
+{"id":"cohere","name":"Cohere","news_24h":0,"news_7d":0,"trending_repos":0,"agent_hits":0,"raw_score":0,"attention_score":0,"rank":10,"top_articles":[]}
+{"id":"perplexity","name":"Perplexity","news_24h":0,"news_7d":0,"trending_repos":0,"agent_hits":0,"raw_score":0,"attention_score":0,"rank":11,"top_articles":[]}
+{"id":"cursor","name":"Cursor","news_24h":0,"news_7d":0,"trending_repos":0,"agent_hits":0,"raw_score":0,"attention_score":0,"rank":12,"top_articles":[]}

2026-10-12/benchmark-registry.jsonl ADDED Viewed

	@@ -0,0 +1,35 @@

+{"id":"mmlu","name":"MMLU","category":"knowledge","description":"Massive Multitask Language Understanding. 57-subject multiple-choice covering law, medicine, history, math, etc. The original general-knowledge benchmark.","released":"2020","size":"15,908 questions","scoreRange":"0-100% accuracy","frontierScore":"~92% (no maintainer board; last traceable figure 2025-08)","frontierModel":null,"frontierDate":null,"frontierSource":null,"status":"saturated","contaminationRisk":"high","maintainer":"Hendrycks et al.","paperUrl":"https://arxiv.org/abs/2009.03300","repoUrl":"https://github.com/hendrycks/test","leaderboardUrl":null,"whoCares":"General assistants. Mostly saturated; use MMLU-Pro for current frontier signal."}
+{"id":"mmlu-pro","name":"MMLU-Pro","category":"knowledge","description":"Harder successor to MMLU. 10-choice questions designed to require multi-step reasoning over memorization. Standard \"is this model smart\" benchmark for general-purpose use.","released":"2024","size":"12,032 questions","scoreRange":"0-100% accuracy","frontierScore":"92.38% (Vals AI run; Vals archived this board as saturated on 2026-09-01 and no longer runs new releases, and the Artificial Analysis board has not added 2026 models)","frontierModel":"Claude Fable 5.1","frontierDate":"2026-09","frontierSource":"https://www.vals.ai/benchmarks/mmlu_pro","status":"active","contaminationRisk":"medium","maintainer":"TIGER Lab","paperUrl":"https://arxiv.org/abs/2406.01574","repoUrl":"https://github.com/TIGER-AI-Lab/MMLU-Pro","leaderboardUrl":"https://huggingface.co/spaces/TIGER-Lab/MMLU-Pro","whoCares":"General chat assistants and research synthesis agents."}
+{"id":"gpqa-diamond","name":"GPQA Diamond","category":"knowledge","description":"Graduate-level Google-Proof Questions, Diamond subset. 198 expert-validated PhD-hard physics, chemistry, biology questions.","released":"2023","size":"198 questions","scoreRange":"0-100% accuracy","frontierScore":"96.3% (Artificial Analysis run)","frontierModel":"GPT-6 Astra (xhigh)","frontierDate":"2026-09","frontierSource":"https://artificialanalysis.ai/evaluations/gpqa-diamond","status":"saturated","contaminationRisk":"low","maintainer":"Rein et al.","paperUrl":"https://arxiv.org/abs/2311.12022","repoUrl":"https://github.com/idavidrein/gpqa","leaderboardUrl":null,"whoCares":"Research agents, technical writing agents, anyone who needs scientific reasoning beyond fact recall. Vals AI archived its GPQA Diamond board as saturated on 2026-09-01, so new releases get no Vals score here."}
+{"id":"hle","name":"Humanity's Last Exam","category":"knowledge","description":"Joint CAIS/Scale benchmark of expert-validated questions across 100+ disciplines, finalized in April 2025 at 2,500 public questions after bug-bounty and searchable items were replaced, with a private held-out set kept for overfitting checks. Vendors report it both with and without tools, and tool setups (search blocklists, fetch restrictions) vary, so check the configuration before comparing.","released":"2025-01","size":"2,500 questions (public set)","scoreRange":"0-100% accuracy","frontierScore":"61.4% no tools (Artificial Analysis run; previous best 59.1%, Claude Fable 5.1)","frontierModel":"Claude Opus 5.5 (adaptive reasoning, max effort, default fallback)","frontierDate":"2026-09","frontierSource":"https://artificialanalysis.ai/evaluations/humanitys-last-exam","status":"active","contaminationRisk":"low","maintainer":"CAIS + Scale AI","paperUrl":"https://lastexam.ai","repoUrl":"https://github.com/centerforaisafety/hle","leaderboardUrl":"https://labs.scale.com/leaderboard/humanitys_last_exam","whoCares":"Frontier-model evaluators. The benchmark every model release in 2026 reports against. The lastexam.ai table is static; the Scale Labs board is the maintained ranking."}
+{"id":"arc-agi-2","name":"ARC-AGI-2","category":"knowledge","description":"Abstract pattern-matching benchmark from Francois Chollet. ARC-AGI-2 is the harder 2025 successor; tests novel problem solving without training-data overlap.","released":"2025-03","size":"~400 tasks public, more private","scoreRange":"0-100% accuracy","frontierScore":"95.0% at $1.12/task (human panel 100%)","frontierModel":"GPT-6 Astra (max)","frontierDate":"2026-09","frontierSource":"https://arcprize.org/results/openai-gpt-6-astra","status":"saturated","contaminationRisk":"low","maintainer":"ARC Prize Foundation","paperUrl":"https://arxiv.org/abs/2412.04604","repoUrl":"https://github.com/arcprize/ARC-AGI-2","leaderboardUrl":"https://arcprize.org/leaderboard","whoCares":"AGI researchers tracking the history of the ARC series. Frontier models now sit within a few points of the human panel, so ARC-AGI-3 carries the open signal."}
+{"id":"math","name":"MATH","category":"math","description":"Competition-level math problems (AMC, AIME, Putnam). Multi-step algebra, geometry, combinatorics. Exact-match scoring.","released":"2021","size":"12,500 problems","scoreRange":"0-100% accuracy","frontierScore":"~99.4% on MATH-500; full Hendrycks set abandoned","frontierModel":null,"frontierDate":null,"frontierSource":null,"status":"saturated","contaminationRisk":"high","maintainer":"Hendrycks et al.","paperUrl":"https://arxiv.org/abs/2103.03874","repoUrl":"https://github.com/hendrycks/math","leaderboardUrl":null,"whoCares":"Quantitative agents. Largely saturated; use AIME 2025 for current signal."}
+{"id":"aime-2025","name":"AIME 2025","category":"math","description":"American Invitational Mathematics Examination 2025. 30 hard high-school competition math problems, integer answers. Released contamination-free in 2025.","released":"2025","size":"30 problems","scoreRange":"0-30 correct","frontierScore":"100.00%","frontierModel":"GPT-5.2 (high and xhigh, tied)","frontierDate":"2026-08","frontierSource":"https://matharena.ai/?comp=aime--aime_2025","status":"deprecated","contaminationRisk":"low","maintainer":"MAA","paperUrl":null,"repoUrl":null,"leaderboardUrl":"https://artificialanalysis.ai","whoCares":"Frontier reasoning evaluators. Standard reasoning-tier test in 2025-2026 model release announcements."}
+{"id":"gsm8k","name":"GSM8K","category":"math","description":"Grade-school math word problems. Multi-step arithmetic with chain-of-thought as the canonical solution form.","released":"2021","size":"8,500 problems","scoreRange":"0-100% accuracy","frontierScore":"99.6 (8-shot)","frontierModel":"MiMo-V2.5-Pro Base","frontierDate":"2026-04","frontierSource":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro","status":"saturated","contaminationRisk":"high","maintainer":"OpenAI","paperUrl":"https://arxiv.org/abs/2110.14168","repoUrl":"https://github.com/openai/grade-school-math","leaderboardUrl":null,"whoCares":"Mostly saturated. Useful as a sanity check that a model can do basic chain-of-thought reasoning."}
+{"id":"humaneval","name":"HumanEval","category":"code","description":"OpenAI's 164 Python programming problems with unit tests. The original code-gen benchmark.","released":"2021","size":"164 problems","scoreRange":"0-100% pass@1","frontierScore":"96.3 pass@1","frontierModel":"o1 Preview and o1 Mini (tied)","frontierDate":"2024-12","frontierSource":"https://evalplus.github.io/leaderboard.html","status":"saturated","contaminationRisk":"high","maintainer":"OpenAI","paperUrl":"https://arxiv.org/abs/2107.03374","repoUrl":"https://github.com/openai/human-eval","leaderboardUrl":"https://evalplus.github.io/leaderboard.html","whoCares":"Capability floor only. Not a frontier signal; use SWE-bench Verified or LiveCodeBench instead."}
+{"id":"mbpp","name":"MBPP","category":"code","description":"Mostly Basic Python Problems. 1k crowd-sourced introductory programming problems.","released":"2021","size":"974 problems","scoreRange":"0-100% pass@1","frontierScore":"95.5 pass@1","frontierModel":"o1 Preview","frontierDate":"2024-12","frontierSource":"https://evalplus.github.io/leaderboard.html","status":"saturated","contaminationRisk":"high","maintainer":"Google Research","paperUrl":"https://arxiv.org/abs/2108.07732","repoUrl":"https://github.com/google-research/google-research/tree/master/mbpp","leaderboardUrl":null,"whoCares":"Same role as HumanEval. Saturated; use newer benchmarks."}
+{"id":"livecodebench","name":"LiveCodeBench","category":"code","description":"Continuously refreshed code benchmark. Pulls competitive programming problems (LeetCode, AtCoder, Codeforces) from after a model's training cutoff. Contamination-free by construction.","released":"2024","size":"1k+ problems, growing","scoreRange":"0-100% pass@1","frontierScore":"87.3 pass@1 (v6, final version)","frontierModel":"o4-mini (high)","frontierDate":"2025-08","frontierSource":"https://livecodebench.github.io/leaderboard.html","status":"saturated","contaminationRisk":"low","maintainer":"UC Berkeley + UW","paperUrl":"https://arxiv.org/abs/2403.07974","repoUrl":"https://github.com/LiveCodeBench/LiveCodeBench","leaderboardUrl":"https://livecodebench.github.io/leaderboard.html","whoCares":"Coding agents. The contamination-resistant alternative to HumanEval. Leaderboard rotates fast as the problem set grows."}
+{"id":"swe-bench-verified","name":"SWE-bench Verified","category":"code","description":"500 human-validated GitHub issues across 12 Python repos. Patch must resolve the issue and pass the project test suite. The official swebench.com board has taken no new Verified entry since February 2026 and still tops out at 79.2% (Claude 4.5 Opus scaffolds), while independent runs and vendor launches now cluster in the 90s, so the spread between harnesses is wider than the spread between models.","released":"2024-08","size":"500 instances","scoreRange":"0-100% resolved","frontierScore":"97.00% (Vals AI bash-only harness, board archived as saturated on 2026-09-01; seven models at 95% or better)","frontierModel":"Claude Opus 5","frontierDate":"2026-09","frontierSource":"https://www.vals.ai/benchmarks/swebench","status":"saturated","contaminationRisk":"medium","maintainer":"OpenAI / Princeton","paperUrl":"https://arxiv.org/abs/2310.06770","repoUrl":"https://github.com/swe-bench/SWE-bench","leaderboardUrl":"https://www.swebench.com/","whoCares":"Coding agents, as a capability floor. Still quoted everywhere, but the top is saturated; use SWE-bench Pro, DeepSWE, or FrontierCode to separate frontier models."}
+{"id":"aider-polyglot","name":"Aider Polyglot","category":"code","description":"225 of the hardest Exercism exercises across C++, Go, Java, JavaScript, Python, Rust. Edit-by-diff scoring (closer to real-world coding work).","released":"2024","size":"225 problems","scoreRange":"0-100% pass@2","frontierScore":"88.0% (diff format)","frontierModel":"GPT-5 (high)","frontierDate":"2025-08","frontierSource":"https://aider.chat/docs/leaderboards/","status":"deprecated","contaminationRisk":"medium","maintainer":"Aider","paperUrl":null,"repoUrl":"https://github.com/Aider-AI/polyglot-benchmark","leaderboardUrl":"https://aider.chat/docs/leaderboards/","whoCares":"Multilingual coding agents. The cross-language complement to SWE-bench."}
+{"id":"terminal-bench","name":"Terminal-Bench","category":"code","description":"Stanford / Anthropic agentic terminal-task benchmark. Each task gives the agent a goal and a sandboxed shell; success measured by deterministic post-condition.","released":"2025-01","size":"~80 tasks","scoreRange":"0-100% solved","frontierScore":"64.5% (1.0, legacy split)","frontierModel":"Apex2 with Claude Sonnet 4.5","frontierDate":"2025-10","frontierSource":"https://tbench.ai/leaderboard/terminal-bench/1.0","status":"deprecated","contaminationRisk":"low","maintainer":"Stanford + Anthropic","paperUrl":"https://arxiv.org/abs/2501.18099","repoUrl":"https://github.com/laude-institute/terminal-bench","leaderboardUrl":"https://www.tbench.ai/","whoCares":"Terminal-shaped agents (Claude Code, Codex CLI, Aider). Tests the loop, not the language model."}
+{"id":"swe-lancer","name":"SWE-Lancer","category":"code","description":"OpenAI benchmark of paid Upwork engineering tasks ($1M+ in real bounties). Includes diff-style fixes and longer feature work judged against original buyer acceptance criteria.","released":"2025-02","size":"1,488 tasks","scoreRange":"0-100% earned","frontierScore":"81.4% (IC Diamond pass rate)","frontierModel":"GPT-5.3-Codex (xhigh)","frontierDate":"2026-02","frontierSource":"https://openai.com/index/introducing-gpt-5-3-codex/","status":"active","contaminationRisk":"low","maintainer":"OpenAI","paperUrl":"https://arxiv.org/abs/2502.12115","repoUrl":"https://github.com/openai/SWELancer-Benchmark","leaderboardUrl":null,"whoCares":"Agentic-coding evaluators. Most realistic eval of \"could you replace a freelance engineer for this task.\""}
+{"id":"mmmu-pro","name":"MMMU-Pro","category":"multimodal","description":"Massive Multi-discipline Multimodal Understanding. College-level questions with images across 30 subjects. Pro version drops text-only solvable variants.","released":"2024","size":"~3,500 questions","scoreRange":"0-100% accuracy","frontierScore":"~86.9% (board aggregates vendor claims, no re-runs)","frontierModel":null,"frontierDate":null,"frontierSource":null,"status":"active","contaminationRisk":"medium","maintainer":"TIGER Lab + IN.AI","paperUrl":"https://arxiv.org/abs/2409.02813","repoUrl":"https://github.com/MMMU-Benchmark/MMMU","leaderboardUrl":"https://mmmu-benchmark.github.io","whoCares":"Vision-capable agents. Standard benchmark in multimodal model release announcements."}
+{"id":"mathvista","name":"MathVista","category":"multimodal","description":"Visual math reasoning across diagrams, charts, geometry, and graphs.","released":"2024","size":"6,141 problems","scoreRange":"0-100% accuracy","frontierScore":"~80%","frontierModel":null,"frontierDate":null,"frontierSource":null,"status":"active","contaminationRisk":"medium","maintainer":"UCLA","paperUrl":"https://arxiv.org/abs/2310.02255","repoUrl":"https://github.com/lupantech/MathVista","leaderboardUrl":"https://mathvista.github.io","whoCares":"Multimodal reasoning agents (chart understanding, scientific diagrams)."}
+{"id":"tau-bench","name":"Tau-Bench","category":"agents","description":"Sierra benchmark of customer-service agents. Tests realistic multi-turn tool-use with deterministic evaluators. Now at tau3: tau2 covers retail, airline, and telecom, while tau3 adds a banking knowledge-retrieval track and a real-time voice track, each ranked separately.","released":"2024-06","size":"~250 tasks","scoreRange":"0-100% pass@1","frontierScore":"87.9% (tau2 text, 3-domain avg pass^1)","frontierModel":"Qwen3.5-397B-A17B (thinking)","frontierDate":"2026-03","frontierSource":"https://taubench.com","status":"active","contaminationRisk":"low","maintainer":"Sierra","paperUrl":"https://arxiv.org/abs/2406.12045","repoUrl":"https://github.com/sierra-research/tau-bench","leaderboardUrl":"https://taubench.com","whoCares":"Customer-service and tool-use agents. The most cited evaluation of \"would this actually work in production support.\""}
+{"id":"gaia","name":"GAIA","category":"agents","description":"General AI Assistants benchmark. 466 real-world questions requiring web browsing, file ops, and multi-step reasoning. Three difficulty levels.","released":"2023","size":"466 tasks","scoreRange":"0-100% accuracy","frontierScore":"93.36% (test split)","frontierModel":"CustomGPT.ai Research Lab v44 (agent scaffold)","frontierDate":"2026-06","frontierSource":"https://huggingface.co/spaces/gaia-benchmark/leaderboard","status":"saturated","contaminationRisk":"low","maintainer":"HF + Meta","paperUrl":"https://arxiv.org/abs/2311.12983","repoUrl":"https://huggingface.co/datasets/gaia-benchmark/GAIA","leaderboardUrl":"https://huggingface.co/spaces/gaia-benchmark/leaderboard","whoCares":"Generalist autonomous agents. Tests the full agent loop, not isolated capabilities. HF Space leaderboard rotates frequently."}
+{"id":"webarena","name":"WebArena","category":"agents","description":"Browser-agent benchmark. 812 tasks across 5 simulated websites (e-commerce, GitLab, Reddit, mapping, content management). Deterministic post-condition checks.","released":"2023","size":"812 tasks","scoreRange":"0-100% solved","frontierScore":"74.3%","frontierModel":"DeepSeek v3.2 with WebTactix","frontierDate":"2026-02","frontierSource":"https://github.com/web-arena-x/webarena","status":"active","contaminationRisk":"low","maintainer":"CMU + UW","paperUrl":"https://arxiv.org/abs/2307.13854","repoUrl":"https://github.com/web-arena-x/webarena","leaderboardUrl":"https://webarena.dev/og/","whoCares":"Browser-automation agents. Standard eval for \"can this agent click around the web reliably.\""}
+{"id":"osworld","name":"OSWorld","category":"agents","description":"Computer-use benchmark. Real Ubuntu, Windows, macOS desktops with 369 tasks covering OS-level workflows (file management, app interaction, research).","released":"2024","size":"369 tasks","scoreRange":"0-100% solved","frontierScore":"90.19% (OSWorld-Verified, 100-step budget); best general model 85.96%","frontierModel":"Intelligence-Indeed Agent","frontierDate":"2026-07","frontierSource":"https://osworld-v1.xlang.ai","status":"saturated","contaminationRisk":"low","maintainer":"HKU + Salesforce","paperUrl":"https://arxiv.org/abs/2404.07972","repoUrl":"https://github.com/xlang-ai/OSWorld","leaderboardUrl":"https://osworld-v1.xlang.ai","whoCares":"Computer-use agents, as a floor check. Near saturation in 2026; OSWorld 2.0 is the long-horizon successor to watch."}
+{"id":"bfcl","name":"BFCL V4","category":"agents","description":"Berkeley Function Calling Leaderboard. Multi-turn, multi-language function calling with parallel and chained tool use.","released":"2024","size":"4,400+ test cases","scoreRange":"0-100% accuracy","frontierScore":"77.47 overall (V4)","frontierModel":"Claude Opus 4.5 (function calling)","frontierDate":"2026-04","frontierSource":"https://gorilla.cs.berkeley.edu/leaderboard.html","status":"active","contaminationRisk":"medium","maintainer":"UC Berkeley","paperUrl":null,"repoUrl":"https://github.com/ShishirPatil/gorilla","leaderboardUrl":"https://gorilla.cs.berkeley.edu/leaderboard.html","whoCares":"Tool-using agents. Most-cited function-calling benchmark; checks parallel and dependent tool calls."}
+{"id":"hcast","name":"HCAST","category":"agents","description":"METR's long-horizon agentic task suite. Measures how long a task an agent can autonomously complete. Calibrated against human researcher time-to-complete.","released":"2024","size":"~70 tasks","scoreRange":"minutes-of-human-work","frontierScore":"50% time horizon of 17.4 hours (above METR's own 16h reliability ceiling)","frontierModel":"Claude Mythos Preview (early, ReAct/Inspect)","frontierDate":"2026-04","frontierSource":"https://metr.org/time-horizons/","status":"active","contaminationRisk":"low","maintainer":"METR","paperUrl":"https://metr.github.io/autonomy-evals-guide/","repoUrl":null,"leaderboardUrl":null,"whoCares":"Frontier autonomy evaluators. The benchmark behind the \"AI does 50% of tasks under N minutes\" framing. METR site is the freshness anchor."}
+{"id":"ruler","name":"RULER","category":"long-context","description":"NVIDIA benchmark for long-context retrieval and reasoning. Multi-needle, variable-tracking, common-words. Tests effective context length, not just claimed.","released":"2024","size":"~13 task categories","scoreRange":"0-100% accuracy","frontierScore":"95.1 at 128K (official table has no 512K or 1M column)","frontierModel":"Jamba-1.5-large","frontierDate":"2024-08","frontierSource":"https://github.com/NVIDIA/RULER","status":"saturated","contaminationRisk":"low","maintainer":"NVIDIA","paperUrl":"https://arxiv.org/abs/2404.06654","repoUrl":"https://github.com/NVIDIA/RULER","leaderboardUrl":null,"whoCares":"Anyone using long-context models. RULER reveals the gap between claimed and effective context length."}
+{"id":"arc-agi-3","name":"ARC-AGI-3","category":"agents","description":"Drops an agent into a novel turn-based game with no instructions, rules, or stated goal. The agent must explore, infer the mechanics, form a goal, and plan toward it. Hundreds of handcrafted environments held privately and reached through an API rather than a static test file.","released":"2026","size":"Hundreds of handcrafted interactive environments","scoreRange":"0-100% environments solved","frontierScore":"62.7% on the standard harness (99.9% with the OpenAI provider adapter harness)","frontierModel":"GPT-6 Astra (max)","frontierDate":"2026-09","frontierSource":"https://arcprize.org/results/openai-gpt-6-astra","status":"active","contaminationRisk":"low","maintainer":"ARC Prize Foundation","paperUrl":"https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf","repoUrl":"https://github.com/arcprize/ARC-AGI-3-Agents","leaderboardUrl":"https://arcprize.org/leaderboard","whoCares":"Anyone betting on agents that must operate in environments nobody documented for them. Check which harness a score used: the provider adapter result is near the ceiling while the standard harness leaves real headroom."}
+{"id":"swe-bench-pro","name":"SWE-bench Pro","category":"code","description":"Long-horizon enterprise software engineering. 1,865 instances across 41 repositories requiring multi-file patches on tasks sized in hours to days. Split three ways into 731 public, 858 held out, and 276 drawn from proprietary startup codebases, so the public-to-private score drop is itself the contamination signal.","released":"2025","size":"1,865 instances across 41 repos","scoreRange":"0-100% resolution rate","frontierScore":"61.50 public split (51.50 private); newer vendor self-reports run far higher and are not on the Scale board","frontierModel":"Muse Spark 1.1","frontierDate":"2026-07","frontierSource":"https://labs.scale.com/leaderboard/swe_bench_pro_public","status":"active","contaminationRisk":"medium","maintainer":"Scale AI (Scale Labs), not the Princeton SWE-bench team","paperUrl":"https://arxiv.org/abs/2509.16941","repoUrl":"https://github.com/scaleapi/SWE-bench_Pro-os","leaderboardUrl":"https://labs.scale.com/leaderboard/swe_bench_pro_public","whoCares":"Teams pointing coding agents at large private codebases rather than tidy open-source issues."}
+{"id":"frontiercode","name":"FrontierCode","category":"code","description":"Scores whether an agent's patch on a real open-source issue is mergeable, not merely test-passing. Grading combines held-out tests with a maintainer-authored rubric covering behavioral correctness, regression safety, test quality, scope discipline, build and lint cleanliness, and project conventions. Roughly 36 maintainers from projects including Celery, Budibase, uppy, and Mattermost authored the tasks. Not to be confused with Epoch AI's FrontierMath.","released":"2026","size":"100 tasks (Main split), 150 (Extended)","scoreRange":"0-100% composite score (the pass rate on task correctness is reported separately and runs higher)","frontierScore":"54.6% score, Main split v1.1, medium effort; Claude Fable 5 (53.5%), Claude Opus 5 (53.4%), and GPT-6 Astra (53.3%) follow within about 1.3 points","frontierModel":"Claude Opus 5.5 (medium)","frontierDate":"2026-09","frontierSource":"https://cognition.com/frontiercode","status":"active","contaminationRisk":"low","maintainer":"Cognition","paperUrl":null,"repoUrl":null,"leaderboardUrl":"https://cognition.com/frontiercode","whoCares":"Anyone shipping agent-written patches into a repo other humans maintain."}
+{"id":"osworld-2","name":"OSWorld 2.0","category":"agents","description":"Long-horizon computer use across 108 workflows in 7 professional domains, taking a skilled human a median 1.6 hours and averaging 318 tool calls. Graded against many weighted checkpoints. Read the metric carefully: binary completion requires every checkpoint to pass, while the partial score averages the fraction satisfied, and vendors usually publish the partial figure. Also check the task release: the official board splits results between v2026.06.24 and v2026.08.08 task files plus an offline no-internet subset, and some vendor figures use modified tasks or grading, so scores from different releases or setups are not comparable.","released":"2026","size":"108 workflows across 7 domains","scoreRange":"0-100% binary completion (partial score reported separately)","frontierScore":"31.43% binary at 500 steps (68.31% partial), v2026.08.08 full set","frontierModel":"Claude Opus 5 (max, batch tool)","frontierDate":"2026-08","frontierSource":"https://osworld-v2.xlang.ai","status":"active","contaminationRisk":"low","maintainer":"XLANG Lab, University of Hong Kong","paperUrl":"https://arxiv.org/abs/2606.29537","repoUrl":"https://github.com/xlang-ai/OSWorld-V2","leaderboardUrl":"https://osworld-v2.xlang.ai","whoCares":"Anyone deploying agents that drive a real desktop instead of an API."}
+{"id":"browsecomp","name":"BrowseComp","category":"agents","description":"1,266 questions whose answers are short and verifiable but buried behind persistent, entangled web research. No official leaderboard exists and every published figure is self-reported, so treat rankings as measuring a full system (model plus tools plus context policy) rather than a bare model.","released":"2025","size":"1,266 questions","scoreRange":"0-100% accuracy","frontierScore":"92.2% (4-agent ultra configuration)","frontierModel":"GPT-5.6 Sol","frontierDate":"2026-07","frontierSource":"https://leaderboard.steel.dev/leaderboards/browsecomp/","status":"active","contaminationRisk":"high","maintainer":"OpenAI","paperUrl":"https://arxiv.org/abs/2504.12516","repoUrl":"https://github.com/openai/simple-evals","leaderboardUrl":"https://leaderboard.steel.dev/leaderboards/browsecomp/","whoCares":"Deep-research and agentic-search products. Scores swing sharply with context-management strategy, so compare like for like."}
+{"id":"gdpval","name":"GDPval","category":"knowledge","description":"Real economically valuable work: 1,320 tasks across 44 occupations and 9 US economic sectors, authored by professionals averaging 14 years of experience. The metric is not accuracy. Occupational experts grade outputs in blind pairwise comparison against human deliverables, producing a win rate, and the widely quoted GDPval-AA variant reports an Elo anchored to a human baseline of 1,000. Never normalize those Elo figures onto a 0-100 scale.","released":"2025","size":"1,320 tasks across 44 occupations","scoreRange":"Win rate vs human deliverables, or Elo anchored at 1,000 (GDPval-AA v2.1)","frontierScore":"1846 Elo (GDPval-AA v2.1, 220 tasks), 111 ahead of Claude Fable 5.1 and 138 ahead of Claude Opus 5; the Elo re-anchors as models join and versions change, so older v2 snapshots do not compare","frontierModel":"Claude Opus 5.5 (adaptive reasoning, max effort, default fallback)","frontierDate":"2026-09","frontierSource":"https://artificialanalysis.ai/evaluations/gdpval-aa","status":"active","contaminationRisk":"low","maintainer":"OpenAI (GDPval-AA v2.1 run by Artificial Analysis)","paperUrl":"https://arxiv.org/abs/2510.04374","repoUrl":null,"leaderboardUrl":"https://artificialanalysis.ai/evaluations/gdpval-aa","whoCares":"Anyone forecasting labor-market impact rather than benchmark scores."}
+{"id":"deepswe","name":"DeepSWE","category":"code","description":"113 original long-horizon engineering tasks across 91 repositories and 5 languages, written from scratch rather than mined from merged pull requests, with hand-written verifiers that test behavior instead of implementation. Prompts run about half the length of SWE-bench Pro's while solutions require roughly 5.5x more code. Note the name collision: an unrelated 2025 open-weight coding agent from Agentica and Together AI is also called DeepSWE.","released":"2026","size":"113 tasks across 91 repos, 5 languages","scoreRange":"0-100% pass rate","frontierScore":"74% (v1.1), a three-way tie within error bars","frontierModel":"GPT-6 Astra (xhigh), Gemini 3.8 Flash (high), and Claude Opus 5 (max), tied","frontierDate":"2026-09","frontierSource":"https://deepswe.datacurve.ai","status":"active","contaminationRisk":"low","maintainer":"Datacurve","paperUrl":"https://arxiv.org/abs/2607.07946","repoUrl":"https://github.com/datacurve-ai/deep-swe","leaderboardUrl":"https://deepswe.datacurve.ai","whoCares":"Teams who suspect their coding agent is pattern-matching scraped commits rather than engineering."}
+{"id":"terminal-bench-2","name":"Terminal-Bench 2.1","category":"agents","description":"89 hard command-line tasks, each in its own containerized environment with a human-written solution and comprehensive tests. Version 2.1 repaired 28 problematic tasks from 2.0 (dependency drift, resource and timeout mismatches, misspecification), and the largest resulting swing was 12.1 points, so 2.0 and 2.1 scores must not be compared directly.","released":"2026","size":"89 tasks","scoreRange":"0-100% tasks solved","frontierScore":"87.27% (Vals AI independent run); vendor self-reports reach 90.6%","frontierModel":"GPT-6 Astra","frontierDate":"2026-09","frontierSource":"https://www.vals.ai/benchmarks/terminal-bench-2-1","status":"saturated","contaminationRisk":"low","maintainer":"Terminal-Bench Team (Harbor harness)","paperUrl":null,"repoUrl":"https://github.com/harbor-framework/terminal-bench-2-1","leaderboardUrl":"https://www.tbench.ai","whoCares":"Still the terminal score most launch tables quote, but Terminal-Bench 4.0 is now the current release and tbench.ai defaults to it."}
+{"id":"terminal-bench-3","name":"Terminal-Bench 3.0","category":"agents","description":"74 tasks across coding, deep learning, finance, engineering, math, and science, including GPU and multi-container environments, assembled by open community contribution behind an automated review pipeline. Superseded in August 2026 by Terminal-Bench 4.0, which pruned and repaired this task set; 3.0 and 4.0 scores are not comparable.","released":"2026","size":"74 tasks across 7 domains","scoreRange":"0-100% tasks solved","frontierScore":"42.7% (+/- 1.6), final 3.0 snapshot","frontierModel":"Claude Opus 5 (max) with mini-SWE-agent","frontierDate":"2026-07","frontierSource":"https://frontierbench.ai","status":"deprecated","contaminationRisk":"low","maintainer":"Terminal-Bench Team (Harbor harness)","paperUrl":null,"repoUrl":null,"leaderboardUrl":"https://www.tbench.ai","whoCares":"Historical reference only. frontierbench.ai now redirects to the Terminal-Bench 4.0 board."}
+{"id":"terminal-bench-4","name":"Terminal-Bench 4.0","category":"agents","description":"The current Terminal-Bench release, now maintained as a continuous benchmark with semantic versioning. Starts from the 3.0 task set, removes 8 tasks (saturated, refusal-prone, publicly solved, or unresolved quality issues), fixes 19 more, and recalibrates resources to a flat 8-hour agent timeout. Every row pairs a model with an agent harness, so it ranks systems rather than bare models. Scores are not comparable with 3.0 or 2.x.","released":"2026-08","size":"66 tasks","scoreRange":"0-100% resolution rate","frontierScore":"65.15% (Vals AI independent run, mini-swe-agent harness, avg@3, board updated 2026-09-29), or 58.08% if the 22 of 198 attempts served by fallback models count as failures; Claude Sonnet 5.5 is one point behind at 64.14%. With its 2026-09-29 update Vals moved the whole board from the Terminus 2 harness to mini-swe-agent, so its earlier figures (Opus 5.5 at 61.62%) are not comparable. Anthropic self-reports 66.4% at xhigh for Opus 5.5 and 70.6% for Sonnet 5.5, and Artificial Analysis measured Opus 5.5 at 59.6%. As of 2026-09-23 the official tbench.ai board had no Opus 5.5 row and was led by GPT-6 Astra (max) with Codex at 58.2% (+/- 2.8)","frontierModel":"Claude Opus 5.5","frontierDate":"2026-09","frontierSource":"https://www.vals.ai/benchmarks/terminal-bench-4","status":"active","contaminationRisk":"low","maintainer":"Terminal-Bench Team (Stanford, Laude Institute; Harbor harness)","paperUrl":null,"repoUrl":"https://github.com/harbor-framework/terminal-bench","leaderboardUrl":"https://www.tbench.ai","whoCares":"Anyone running agents with a real shell. The version Anthropic, OpenAI, Google, and DeepSeek all cited in their September 2026 launches."}
+{"id":"automationbench","name":"AutomationBench","category":"agents","description":"Zapier benchmark of cross-application business workflows over REST APIs across 47 simulated SaaS apps in sales, marketing, operations, support, finance, and HR. Agents must discover endpoints, follow policies buried in the environment, and avoid decoy records. Grading is deterministic end-state assertions with no LLM judge, strict pass or fail per task. The official board scores a held-out private set; the public 600-task repo set runs easier, so local numbers do not match it one to one.","released":"2026-04","size":"600+ held-out tasks across 6 domains (public set of 600 released separately)","scoreRange":"0-100% tasks completed correctly","frontierScore":"42.47% (v1.0.6 private set); GPT-6 Astra (max) is close behind at 41.4%","frontierModel":"Claude Opus 5.5 (max, default fallbacks)","frontierDate":"2026-09","frontierSource":"https://zapier.com/benchmarks","status":"active","contaminationRisk":"low","maintainer":"Zapier","paperUrl":"https://arxiv.org/abs/2604.18934","repoUrl":"https://github.com/zapier/AutomationBench","leaderboardUrl":"https://zapier.com/benchmarks","whoCares":"Teams deploying agents into back-office workflows. Watch for fallback configurations: the Claude Fable 5.1 row lets Claude Opus 5 finish steps the Fable safety classifier refuses, and the Claude Opus 5.5 rows rerun refused tasks with Anthropic's default fallback routing."}

2026-10-12/benchmarks.jsonl ADDED Viewed

	@@ -0,0 +1,10 @@

+{"id":"swe_bench","name":"SWE-bench","description":"Real-world software engineering tasks from GitHub issues (SWE-bench Verified)","maxScore":100}
+{"id":"mmlu_pro","name":"MMLU-Pro","description":"General knowledge and reasoning across 57 subjects","maxScore":100}
+{"id":"human_eval","name":"HumanEval","description":"Python code generation and problem solving","maxScore":100}
+{"id":"gpqa_diamond","name":"GPQA Diamond","description":"Graduate-level science questions verified by domain experts","maxScore":100}
+{"id":"math","name":"MATH","description":"Competition-level mathematics problems","maxScore":100}
+{"id":"osworld_2","name":"OSWorld 2.0","description":"Computer use across real desktop applications and multi-step GUI tasks","maxScore":100}
+{"id":"browsecomp","name":"BrowseComp","description":"Agentic web search and browsing over hard-to-find facts","maxScore":100}
+{"id":"frontier_code","name":"FrontierCode v1.1","description":"Agentic coding on frontier software engineering tasks (Main split)","maxScore":100}
+{"id":"terminal_bench_4","name":"Terminal-Bench 4.0","description":"Long-horizon agentic work in a terminal across software, science, ML, operations, hardware, security, and media (66 tasks, all-or-nothing verifiers)","maxScore":100}
+{"id":"hle_tools","name":"Humanity's Last Exam (tools)","description":"Multidisciplinary expert-level reasoning with tool access","maxScore":100}

2026-10-12/compute-providers.jsonl ADDED Viewed

	@@ -0,0 +1,18 @@

+{"id":"lambda","name":"Lambda","vendor":"Lambda","type":"gpu-cloud","gpus":["B200","H100","A100","GH200","A6000"],"pricingModel":"On-demand instances + 1-Click Clusters (2 weeks to 1 year) + reserved","startingPrice":"H100 from $3.29/hr on-demand (1x PCIe); 8x H100 SXM $3.99/GPU/hr","onDemand":true,"spotPricing":false,"regions":"US (multiple), EU, APAC","aiServices":["Lambda Stack (preinstalled CUDA + frameworks)","1-Click Clusters","managed Kubernetes","private cloud"],"bestFor":"Researchers + AI startups wanting bare-metal GPU without hyperscaler complexity","url":"https://lambda.ai","notes":"Founded 2012. The default 'GPU cloud' for indie AI research and startup deployments. Strong CLI experience. Now at lambda.ai (lambdalabs.com redirects)."}
+{"id":"coreweave","name":"CoreWeave","vendor":"CoreWeave","type":"gpu-cloud","gpus":["GB300 NVL72","GB200 NVL72","B300","B200","H200","H100","RTX PRO 6000","GH200","L40S","A100"],"pricingModel":"On-demand + spot + reserved capacity","startingPrice":"HGX H100 (8x) $49.24/hr on-demand; $19.71/hr spot","onDemand":true,"spotPricing":true,"regions":"North America, Europe","aiServices":["Kubernetes-first","managed inference","SUNK scheduler (Slurm-on-K8s)","NVLink fabric"],"bestFor":"Frontier labs + enterprise AI training; second-largest H100/H200 fleet after AWS","url":"https://www.coreweave.com","notes":"IPO'd 2025. Anchor tenant for many frontier lab compute deals. Specializes in NVIDIA-only large-scale training; reported 1.5 GW active power in Q2 2026 and the first Vera Rubin NVL72 bring-up."}
+{"id":"crusoe","name":"Crusoe","vendor":"Crusoe","type":"gpu-cloud","gpus":["GB200","B200","H200","H100","A100","L40S","MI355X","MI300X"],"pricingModel":"On-demand + spot + reserved","startingPrice":"H100 from $3.90/GPU-hr on-demand","onDemand":true,"spotPricing":true,"regions":"US (TX, IA, OK, ND); building Abilene Texas mega-site","aiServices":["AI-first data centers","managed inference","GPU clusters","sustainable energy sourcing"],"bestFor":"Enterprise AI training where energy story matters; mega-cluster customers","url":"https://www.crusoe.ai","notes":"Founded as 'Crusoe Energy' (stranded methane); pivoted to AI-first data centers. Builds the Abilene Texas Stargate campus. Rents both NVIDIA and AMD (MI355X, MI300X) GPUs."}
+{"id":"nebius","name":"Nebius","vendor":"Nebius (formerly Yandex N.V.)","type":"gpu-cloud","gpus":["GB300","GB200","B300","B200","H200","H100","RTX PRO 6000","L40S"],"pricingModel":"On-demand + reserved + preemptible","startingPrice":"H100 $3.85/GPU-hr on-demand; $2.15 preemptible","onDemand":true,"spotPricing":true,"regions":"EU (Finland, France), UK, Israel, Iceland, US (Kansas City, Woodbury MN)","aiServices":["vLLM serving","fine-tuning service","managed K8s","NVLink fabric"],"bestFor":"EU customers needing data residency; teams wanting preemptible H100 access","url":"https://nebius.com","notes":"Reorganized after Yandex split. EU-headquartered. Lists GB300, B300, and B200 capacity alongside H100/H200 in 2026; NVIDIA says Nebius will be first for Groq 3 LPX volume shipments."}
+{"id":"together-gpu-clusters","name":"Together GPU Clusters","vendor":"Together AI","type":"gpu-cloud","gpus":["GB300 NVL72","GB200 NVL72","B300","B200","H200","H100"],"pricingModel":"Preemptible + on-demand + reserved (7 to 180 days, longer by quote)","startingPrice":"HGX H100 $3.99/GPU-hr on-demand; $1.99 preemptible; from $3.19 reserved (91 to 180 days)","onDemand":true,"spotPricing":true,"regions":"Not listed on pricing page","aiServices":["serverless inference","dedicated inference","fine-tuning","code sandbox","managed storage"],"bestFor":"Teams that want training clusters and open-model inference from one vendor","url":"https://www.together.ai/pricing","notes":"Best known for serverless open-model inference; GPU Clusters add hourly, preemptible, and reserved NVIDIA capacity up to GB300 NVL72."}
+{"id":"vast-ai","name":"Vast.ai","vendor":"Vast.ai","type":"marketplace","gpus":["H100","A100","RTX 4090","RTX 3090","RTX 6000","consumer GPUs"],"pricingModel":"Marketplace (host bid + offer)","startingPrice":"RTX 3090 from $0.20/hr; H100 from $1.50/hr (varies wildly)","onDemand":true,"spotPricing":true,"regions":"Global (host-dependent)","aiServices":["Docker-only deployments","SSH access","image library"],"bestFor":"Hobbyist + research workloads; cheapest path for short experiments","url":"https://vast.ai","notes":"Marketplace of independent GPU hosts. Heterogeneous quality; verified hosts available. The cheapest GPU rentals in the catalog."}
+{"id":"runpod","name":"RunPod","vendor":"RunPod","type":"marketplace","gpus":["B300","B200","H200","H100","A100","RTX PRO 6000","L40S","RTX 5090","RTX 4090"],"pricingModel":"Secure Cloud (RunPod-managed) + Community Cloud (peer hosts) + Serverless + Clusters","startingPrice":"H100 PCIe $2.89/hr, H100 SXM $3.29/hr (Pods)","onDemand":true,"spotPricing":true,"regions":"Global (30+ regions)","aiServices":["Serverless GPU functions","instant deploy templates","persistent volumes","instant clusters"],"bestFor":"AI app developers wanting per-second serverless GPU + on-demand instances","url":"https://www.runpod.io","notes":"Strong serverless GPU offering. Fastest cold-start in the marketplace category."}
+{"id":"paperspace","name":"DigitalOcean GPU Droplets (formerly Paperspace)","vendor":"DigitalOcean","type":"gpu-cloud","gpus":["B300","H200","H100","MI355X","MI350X","MI325X","MI300X","L40S","RTX 4000 Ada"],"pricingModel":"On-demand (per-second billing) + 12-month reserved + spot","startingPrice":"H100 $4.41/GPU/hr on-demand; $3.26 12-month reserved","onDemand":true,"spotPricing":true,"regions":"US (multiple), EU","aiServices":["GPU Droplets","spot GPU Droplets","12-month reserved capacity"],"bestFor":"Small teams already on DigitalOcean wanting simple GPU VMs with NVIDIA or AMD options","url":"https://www.digitalocean.com/pricing/gpu-droplets","notes":"Paperspace was acquired by DigitalOcean in 2023 and its site now points users to DigitalOcean. GPU Droplet pricing changed August 1 2026; lists AMD MI300X, MI325X, MI350X, and MI355X next to NVIDIA."}
+{"id":"aws-ec2-gpu","name":"AWS EC2 GPU","vendor":"Amazon Web Services","type":"hyperscaler","gpus":["GB300 (P6e-GB300)","GB200 (P6e-GB200)","B300 (P6-B300)","B200 (P6-B200)","H200 (P5e)","H100 (P5)","A100 (P4d)","V100 (P3)"],"pricingModel":"On-demand + Reserved + Spot + Capacity Reservations + Trainium dedicated","startingPrice":"P5 (8x H100) $55.04/hr on-demand in us-east-1; P6-B200 $113.93/hr","onDemand":true,"spotPricing":true,"regions":"Global (32 regions)","aiServices":["Bedrock (managed model APIs)","SageMaker","Trainium / Inferentia","Capacity Blocks","EFA networking"],"bestFor":"Enterprise + regulated industries needing AWS ecosystem integration","url":"https://aws.amazon.com/ec2/instance-types/p6/","notes":"Largest cloud AI compute footprint. Bedrock is the managed-model layer; Trainium is AWS's in-house silicon (Anthropic uses these). P6e UltraServers bring GB200 and GB300 NVL72 racks."}
+{"id":"azure-ml","name":"Azure ML / OpenAI Service","vendor":"Microsoft","type":"hyperscaler","gpus":["GB200 (ND GB200 v6)","H200 (NDv5)","H100 (ND H100 v5)","A100 (ND A100)"],"pricingModel":"On-demand + Reserved + Spot + Azure Reserved Instances","startingPrice":"ND96isr H100 v5 $98.32/hr; ND GB200 v6 $108.16/hr (East US, Linux)","onDemand":true,"spotPricing":true,"regions":"Global (60+ regions)","aiServices":["Azure OpenAI Service (private GPT)","Azure AI Foundry","Phi-4 hosting","Maia 200 (in-house inference silicon)"],"bestFor":"Enterprise customers anchored on Microsoft ecosystem; private GPT deployments","url":"https://azure.microsoft.com/en-us/products/machine-learning/","notes":"Strongest first-party OpenAI hosting (Azure OpenAI Service). Maia 200 is Microsoft's in-house inference silicon, deployed in Azure US Central since January 2026, paralleling Trainium / TPU."}
+{"id":"gcp-vertex","name":"GCP Vertex AI / Compute Engine GPU","vendor":"Google","type":"hyperscaler","gpus":["B200 (A4)","H200 (A3 Ultra)","H100 (A3)","A100","L4","TPU7x Ironwood","TPU v6e Trillium","TPU v5p","TPU v5e"],"pricingModel":"On-demand + Sustained Use + Committed Use Discounts + Spot + DWS Flex-start + TPU chip-hours","startingPrice":"A3 High (8x H100) $88.49/hr; TPU v5p $4.20/chip-hour, Trillium $2.70, Ironwood $12.00 (US regions)","onDemand":true,"spotPricing":true,"regions":"Global (40+ regions)","aiServices":["Vertex AI","Gemini API","Ironwood and Trillium TPU pods","Model Garden","AutoML"],"bestFor":"Customers needing TPU access; Gemini-anchored stacks","url":"https://cloud.google.com/vertex-ai","notes":"Only major cloud with TPU access. Vertex AI is the managed-model layer. TPU pods scale to 9,216 chips on Ironwood, which went GA in March 2026."}
+{"id":"oci-gpu","name":"Oracle Cloud Infrastructure","vendor":"Oracle","type":"hyperscaler","gpus":["GB300","GB200","B300","B200","H200","H100","A100","L40S","MI355X","MI300X"],"pricingModel":"On-demand + Annual flex commitments + Bare metal","startingPrice":"H100 $10.00/GPU-hr ($80.00/hr for BM.GPU.H100.8)","onDemand":true,"spotPricing":false,"regions":"Global (50+ regions, OCI Generation 2)","aiServices":["OCI Generative AI","Cohere partnership (private deployments)","GPU Bare Metal"],"bestFor":"Enterprise + government wanting NVIDIA cluster scale at competitive bare-metal pricing","url":"https://www.oracle.com/cloud/compute/gpu/","notes":"Won large frontier-lab contracts (OpenAI Stargate, xAI, Meta) on bare-metal GPU pricing. Price list covers GB300 at $18.00 and AMD MI355X at $8.60 per GPU-hour."}
+{"id":"modal","name":"Modal","vendor":"Modal Labs","type":"ai-serverless","gpus":["B300","B200","H200","H100","RTX PRO 6000","A100","L40S","A10","L4","T4"],"pricingModel":"Pay-per-second compute + $30/mo free credit","startingPrice":"H100 from $3.95/hr; per-second billing","onDemand":true,"spotPricing":false,"regions":"US (multi-region)","aiServices":["serverless functions","instant cold-start GPU","volumes","web endpoints","cron jobs","distributed training"],"bestFor":"Python developers wanting serverless GPU without container ops","url":"https://modal.com","notes":"Per-second GPU billing with sub-second cold-start. Best Python developer ergonomics in the AI-serverless category."}
+{"id":"replicate-compute","name":"Replicate","vendor":"Replicate (Cloudflare)","type":"ai-serverless","gpus":["H100","A100","A40","T4","CPU"],"pricingModel":"Pay-per-second on the model you run","startingPrice":"H100 from $0.001525/sec ($5.49/hr)","onDemand":true,"spotPricing":false,"regions":"US, EU","aiServices":["model marketplace (thousands of public models)","API endpoints","webhooks","fine-tuning"],"bestFor":"Image / video / voice model developers wanting one-API model serving","url":"https://replicate.com","notes":"Pay-per-second model serving with a large public model catalog. Replicate joined Cloudflare (announced December 1 2025) and keeps its brand. Strong fit for media-generation pipelines."}
+{"id":"beam","name":"Beam","vendor":"Beam.cloud","type":"ai-serverless","gpus":["B200","H200","H100","RTX PRO 6000","A100","L40S","RTX 5090","A6000","RTX 4090"],"pricingModel":"Serverless (billed per millisecond) + flat-rate on-demand machines + reserved clusters","startingPrice":"H100 on-demand from $1.74/hr; serverless $0.000986/sec (about $3.55/hr)","onDemand":true,"spotPricing":false,"regions":"US","aiServices":["serverless GPU functions","on-demand GPU machines","sandboxes","bring your own cloud"],"bestFor":"Teams wanting cheap on-demand H100 machines and serverless GPU from one SDK","url":"https://www.beam.cloud","notes":"Modal competitor that also sells flat-rate on-demand machines and InfiniBand clusters. $30 free credit every month."}
+{"id":"cerebras-cloud","name":"Cerebras Cloud","vendor":"Cerebras","type":"specialized","gpus":["CS-3 (WSE-3 wafer)"],"pricingModel":"Per-token API (proprietary inference)","startingPrice":"gpt-oss-120b at $0.35 input / $0.75 output per 1M","onDemand":true,"spotPricing":false,"regions":"US","aiServices":["ultra-fast inference","training service","Cerebras Inference Cloud"],"bestFor":"Inference workloads needing maximum tokens/sec on open-weight models","url":"https://www.cerebras.ai/inference","notes":"Unique wafer-scale silicon. Cerebras lists about 3,000 tokens/sec on gpt-oss-120b."}
+{"id":"sambanova","name":"SambaNova Cloud","vendor":"SambaNova","type":"specialized","gpus":["SN40L (in-house RDU)"],"pricingModel":"Per-token API + dedicated capacity","startingPrice":"Llama 3.3 70B at $0.60 input / $1.20 output per 1M","onDemand":true,"spotPricing":false,"regions":"US","aiServices":["fast inference","fine-tuning service","private deployments"],"bestFor":"Enterprise inference; on-prem private cloud option","url":"https://sambanova.ai/products/sambacloud","notes":"Custom RDU silicon. Strong on-prem private deployment story. Less hype than Cerebras but solid customer list."}
+{"id":"fireworks-compute","name":"Fireworks Compute","vendor":"Fireworks AI","type":"ai-serverless","gpus":["GB300","B300","B200","H200","H100"],"pricingModel":"Per-token (serverless) + on-demand deployments (billed per GPU-second)","startingPrice":"H100 on-demand deployment $8.00/GPU-hr (from Sep 1 2026)","onDemand":true,"spotPricing":false,"regions":"US, EU","aiServices":["serverless model APIs","on-demand deployments","fine-tuning","multi-LoRA serving"],"bestFor":"Production inference on open-weights models with dedicated deployment option","url":"https://fireworks.ai","notes":"Inference-first platform. Multi-LoRA serving (many fine-tunes share one base GPU) is the pricing differentiator."}

2026-10-12/conferences.jsonl ADDED Viewed

	@@ -0,0 +1,28 @@

+{"id":"aaai-2026","name":"AAAI 2026","category":"research","startDate":"2026-01-20","endDate":"2026-01-27","city":"Singapore","country":"Singapore","format":"in-person","paperDeadline":"2025-08-01","registrationOpen":false,"themes":["AI broadly","agents","planning","reasoning"],"url":"https://aaai.org/conference/aaai/aaai-26/","notes":"AAAI Conference on Artificial Intelligence. Broadest AI venue (covers more than just deep learning). 2026 edition held at Singapore EXPO. Big presence from China-based research."}
+{"id":"gtc-2026","name":"NVIDIA GTC 2026","category":"industry","startDate":"2026-03-16","endDate":"2026-03-19","city":"San Jose, CA","country":"USA","format":"hybrid","paperDeadline":null,"registrationOpen":false,"themes":["Blackwell","AI factories","Omniverse","robotics"],"url":"https://www.nvidia.com/gtc/","notes":"NVIDIA's GPU Technology Conference. Major chip + AI infrastructure announcements; Jensen keynote is the AI hardware industry's State of the Union."}
+{"id":"iclr-2026","name":"ICLR 2026","category":"research","startDate":"2026-04-23","endDate":"2026-04-27","city":"Rio de Janeiro","country":"Brazil","format":"hybrid","paperDeadline":"2025-09-24","registrationOpen":false,"themes":["representation learning","foundation models","agents","safety"],"url":"https://iclr.cc/Conferences/2026","notes":"International Conference on Learning Representations. Top deep-learning venue along with NeurIPS and ICML. 2026 edition held at Riocentro in Rio de Janeiro; main conference April 23 to 25, workshops April 26 and 27."}
+{"id":"code-with-claude-2026","name":"Code with Claude 2026","category":"industry","startDate":"2026-05-06","endDate":"2026-05-06","city":"San Francisco, CA","country":"USA","format":"hybrid","paperDeadline":null,"registrationOpen":false,"themes":["Claude Code","MCP","agent SDK","Claude API"],"url":"https://claude.com/code-with-claude","notes":"Anthropic's developer conference. San Francisco edition May 6, followed by London (May 19) and Tokyo (June 10), all livestreamed. Free to attend; in-person seats allocated by application."}
+{"id":"google-io-2026","name":"Google I/O 2026","category":"industry","startDate":"2026-05-19","endDate":"2026-05-20","city":"Mountain View, CA","country":"USA","format":"hybrid","paperDeadline":null,"registrationOpen":false,"themes":["Gemini","Android","Workspace AI","developer tools"],"url":"https://io.google/2026/","notes":"Google's annual developer conference, held at Shoreline Amphitheatre and online. Major Gemini announcement venue."}
+{"id":"microsoft-build-2026","name":"Microsoft Build 2026","category":"industry","startDate":"2026-06-02","endDate":"2026-06-03","city":"San Francisco, CA","country":"USA","format":"hybrid","paperDeadline":null,"registrationOpen":false,"themes":["Copilot","Azure AI","GitHub Copilot","agents"],"url":"https://build.microsoft.com","notes":"Microsoft's developer conference, held in San Francisco with free online registration. Copilot ecosystem + Azure AI announcements."}
+{"id":"cvpr-2026","name":"CVPR 2026","category":"research","startDate":"2026-06-03","endDate":"2026-06-07","city":"Denver, CO","country":"USA","format":"hybrid","paperDeadline":"2025-11-13","registrationOpen":false,"themes":["computer vision","multimodal","generative","3D"],"url":"https://cvpr.thecvf.com/Conferences/2026","notes":"Conference on Computer Vision and Pattern Recognition. Top vision venue. 2026 edition held at the Colorado Convention Center, with virtual registration and recorded sessions."}
+{"id":"wwdc-2026","name":"Apple WWDC 2026","category":"industry","startDate":"2026-06-08","endDate":"2026-06-12","city":"Cupertino, CA","country":"USA","format":"hybrid","paperDeadline":null,"registrationOpen":false,"themes":["Apple Intelligence","iOS","macOS","on-device AI"],"url":"https://developer.apple.com/wwdc26/","notes":"Apple's developer conference, held online with an in-person special event at Apple Park on June 8. On-device AI announcements + Apple Intelligence updates."}
+{"id":"cursor-compile-2026","name":"Cursor Compile 2026","category":"developer","startDate":"2026-06-16","endDate":"2026-06-16","city":"San Francisco, CA","country":"USA","format":"in-person","paperDeadline":null,"registrationOpen":false,"themes":["Cursor agents","AI coding","future of software"],"url":"https://cursor.com/compile","notes":"Cursor's first conference, a one-day event in San Francisco. Compile on the Road stops in London (Sept 16), Amsterdam (Sept 22), Sydney (Oct 14), and New York (Nov 5), all sold out."}
+{"id":"aiengineer-summit-2026","name":"AI Engineer World's Fair 2026","category":"developer","startDate":"2026-06-29","endDate":"2026-07-02","city":"San Francisco, CA","country":"USA","format":"in-person","paperDeadline":null,"registrationOpen":false,"themes":["production AI","agents","evals","RAG"],"url":"https://www.ai.engineer/worldsfair/2026","notes":"Largest practitioner-focused AI engineering conference. swyx-organized; standard ground-truth check on what's working in production."}
+{"id":"acl-2026","name":"ACL 2026","category":"research","startDate":"2026-07-02","endDate":"2026-07-07","city":"San Diego, CA","country":"USA","format":"hybrid","paperDeadline":"2026-01-05","registrationOpen":false,"themes":["NLP","multilingual","computational linguistics","discourse"],"url":"https://2026.aclweb.org","notes":"Annual Meeting of the Association for Computational Linguistics. The traditional NLP venue; LLM submissions dominate post-2022. 2026 edition: tutorials July 2, workshops July 3 and 4, main conference July 5 to 7."}
+{"id":"icml-2026","name":"ICML 2026","category":"research","startDate":"2026-07-06","endDate":"2026-07-11","city":"Seoul","country":"South Korea","format":"hybrid","paperDeadline":"2026-01-28","registrationOpen":false,"themes":["ML theory","foundation models","reinforcement learning","fairness"],"url":"https://icml.cc/Conferences/2026","notes":"International Conference on Machine Learning. Co-flagship with NeurIPS. 2026 edition held at COEX in Seoul; tutorials July 6, main conference July 7 to 9, workshops July 10 and 11."}
+{"id":"openai-devday-2026","name":"OpenAI DevDay 2026","category":"industry","startDate":"2026-09-29","endDate":"2026-09-29","city":"San Francisco, CA","country":"USA","format":"hybrid","paperDeadline":null,"registrationOpen":false,"themes":["GPT","Apps SDK","agents","Codex"],"url":"https://devday.openai.com","notes":"OpenAI's annual developer conference at Fort Mason. In-person applications are closed; the opening keynote is livestreamed. DevDay Exchanges run in Bengaluru, Tokyo, Seoul, Paris, Berlin, London, Sao Paulo, and Mexico City."}
+{"id":"colm-2026","name":"COLM 2026","category":"research","startDate":"2026-10-06","endDate":"2026-10-09","city":"San Francisco, CA","country":"USA","format":"in-person","paperDeadline":"2026-03-31","registrationOpen":false,"themes":["language models","evaluation","alignment","multilingual"],"url":"https://colm.cc","notes":"Conference on Language Modeling. Newer venue (founded 2024) focused specifically on LLMs. 2026 edition at Hilton Union Square; main conference Oct 6 to 8, workshops Oct 9. General registration is sold out (waitlist only)."}
+{"id":"emnlp-2026","name":"EMNLP 2026","category":"research","startDate":"2026-10-24","endDate":"2026-10-29","city":"Budapest","country":"Hungary","format":"hybrid","paperDeadline":"2026-05-25","registrationOpen":true,"themes":["empirical methods in NLP","evaluation","multilingual"],"url":"https://2026.emnlp.org","notes":"Empirical Methods in Natural Language Processing. Companion to ACL. In-person and virtual presentation options; workshops Oct 28 and 29. Early registration closes Sept 24."}
+{"id":"corl-2026","name":"CoRL 2026","category":"research","startDate":"2026-11-09","endDate":"2026-11-12","city":"Austin, TX","country":"USA","format":"in-person","paperDeadline":"2026-05-28","registrationOpen":true,"themes":["robot learning","robot foundation models","imitation learning","reinforcement learning"],"url":"https://www.corl.org","notes":"Conference on Robot Learning. Selective single-track venue at the intersection of robotics and machine learning. Main conference Nov 9 to 11, workshops Nov 12, across JW Marriott Austin, Hilton Austin, and Bass Hall."}
+{"id":"aiengineer-code-2026","name":"AI Engineer Code Summit 2026","category":"developer","startDate":"2026-11-10","endDate":"2026-11-12","city":"San Francisco, CA","country":"USA","format":"in-person","paperDeadline":null,"registrationOpen":true,"themes":["coding agents","developer tools","production workflows"],"url":"https://www.ai.engineer/code/2026","notes":"Application-only, single-track AI Engineer conference at Hilton San Francisco Union Square. Speaker applications close Oct 11, 2026."}
+{"id":"aws-reinvent-2026","name":"AWS re:Invent 2026","category":"industry","startDate":"2026-11-30","endDate":"2026-12-04","city":"Las Vegas, NV","country":"USA","format":"hybrid","paperDeadline":null,"registrationOpen":true,"themes":["Bedrock","Trainium","AI infrastructure","enterprise AI"],"url":"https://aws.amazon.com/events/reinvent","notes":"AWS's annual customer conference. Major Bedrock + Trainium announcements; Anthropic typically appears in keynotes. Registration open with early bird pricing."}
+{"id":"gtc-dc-2026","name":"NVIDIA GTC Washington, D.C. 2026","category":"industry","startDate":"2026-11-30","endDate":"2026-12-03","city":"Washington, D.C.","country":"USA","format":"in-person","paperDeadline":null,"registrationOpen":false,"themes":["AI factories","open models","physical AI","HPC"],"url":"https://www.nvidia.com/gtc/dc/","notes":"NVIDIA's policy and public-sector GTC at the Ronald Reagan Building and International Trade Center, with a Jensen Huang keynote. Registration not yet open."}
+{"id":"neurips-2026","name":"NeurIPS 2026","category":"research","startDate":"2026-12-06","endDate":"2026-12-12","city":"Sydney","country":"Australia","format":"hybrid","paperDeadline":"2026-05-06","registrationOpen":true,"themes":["neural networks","foundation models","AGI","datasets","safety"],"url":"https://neurips.cc/Conferences/2026","notes":"Largest AI/ML research conference. 2026 main meeting is in Sydney (main conference Dec 8 to 10), with satellite meetings in Atlanta and Paris running Dec 9 to 13. Virtual-only pass available; early registration closes Oct 30. Datasets + Benchmarks track is the canonical venue for new evaluations."}
+{"id":"aaai-2027","name":"AAAI 2027","category":"research","startDate":"2027-02-16","endDate":"2027-02-23","city":"Montreal","country":"Canada","format":"in-person","paperDeadline":"2026-07-28","registrationOpen":false,"themes":["AI broadly","agents","planning","reasoning"],"url":"https://aaai.org/conference/aaai/aaai-27/","notes":"AAAI Conference on Artificial Intelligence. 2027 edition at the Palais des Congres de Montreal. Main track abstracts were due July 21, 2026."}
+{"id":"eacl-2027","name":"EACL 2027","category":"research","startDate":"2027-03-09","endDate":"2027-03-14","city":"Athens","country":"Greece","format":"in-person","paperDeadline":"2026-08-03","registrationOpen":false,"themes":["NLP","multilingual","computational linguistics"],"url":"https://2027.eacl.org","notes":"Conference of the European Chapter of the ACL. Papers submitted through ARR; specific venue not yet announced."}
+{"id":"gtc-2027","name":"NVIDIA GTC 2027","category":"industry","startDate":"2027-03-15","endDate":"2027-03-18","city":"San Jose, CA","country":"USA","format":"in-person","paperDeadline":null,"registrationOpen":false,"themes":["AI factories","agentic AI","inference","physical AI"],"url":"https://www.nvidia.com/gtc/","notes":"NVIDIA's flagship GPU Technology Conference returns to San Jose. Registration not yet open."}
+{"id":"iclr-2027","name":"ICLR 2027","category":"research","startDate":"2027-04-26","endDate":"2027-04-30","city":"California (city TBA)","country":"USA","format":"in-person","paperDeadline":"2026-09-25","registrationOpen":false,"themes":["representation learning","foundation models","agents","safety"],"url":"https://iclr.cc/Conferences/2027","notes":"International Conference on Learning Representations. Official site lists only California so far; its header image shows Moscone Center in San Francisco, but the venue city is not yet announced. Abstracts due Sept 18, 2026. Days four and five are workshops."}
+{"id":"naacl-2027","name":"NAACL 2027","category":"research","startDate":"2027-06-01","endDate":"2027-06-05","city":"San Francisco, CA","country":"USA","format":"in-person","paperDeadline":"2026-10-12","registrationOpen":false,"themes":["NLP","language models","evaluation"],"url":"https://2027.naacl.org","notes":"Annual Conference of the Nations of the Americas Chapter of the ACL. ARR submission deadline Oct 12, 2026 per the official site, with details still pending."}
+{"id":"cvpr-2027","name":"CVPR 2027","category":"research","startDate":"2027-06-20","endDate":"2027-06-25","city":"Seattle, WA","country":"USA","format":"in-person","paperDeadline":"2026-11-16","registrationOpen":false,"themes":["computer vision","multimodal","generative","3D"],"url":"https://cvpr.thecvf.com/Conferences/2027","notes":"Conference on Computer Vision and Pattern Recognition. Workshops and tutorials June 20 and 21, main conference June 22 to 25. Paper registration closes Nov 10, 2026."}
+{"id":"aiengineer-worldsfair-2027","name":"AI Engineer World's Fair 2027","category":"developer","startDate":"2027-06-29","endDate":"2027-07-02","city":"San Francisco, CA","country":"USA","format":"in-person","paperDeadline":null,"registrationOpen":false,"themes":["production AI","agents","evals","coding agents"],"url":"https://www.ai.engineer/worldsfair/2027","notes":"Returns to Moscone West, targeting 7,000+ AI engineers, founders, and VPs of AI. Program and tickets not yet announced."}
+{"id":"acl-2027","name":"ACL 2027","category":"research","startDate":"2027-08-17","endDate":"2027-08-22","city":"Kyoto","country":"Japan","format":"in-person","paperDeadline":null,"registrationOpen":false,"themes":["NLP","multilingual","computational linguistics"],"url":"https://2027.aclweb.org","notes":"65th Annual Meeting of the Association for Computational Linguistics. ARR submission deadline not yet announced."}

2026-10-12/embeddings.jsonl ADDED Viewed

	@@ -0,0 +1,34 @@

+{"id":"text-embedding-3-large","name":"text-embedding-3-large","provider":"OpenAI","type":"embedding","dimensions":3072,"maxInputTokens":8192,"pricePer1MTokens":0.13,"pricingNote":"$0.13 per 1M input tokens. Reducible-dimensions: pass `dimensions` to truncate to 256/1024/etc with minor quality loss.","openSource":false,"license":"Proprietary","released":"2024-01-25","notes":"OpenAI flagship embedding and still its newest; no successor has shipped. Strong on English retrieval; supports Matryoshka truncation to lower dimensions for cheaper storage. Azure lists a 2028-02-09 retirement for its hosted copy.","multilingual":true,"url":"https://developers.openai.com/api/docs/guides/embeddings","mtebAvg":64.6}
+{"id":"text-embedding-3-small","name":"text-embedding-3-small","provider":"OpenAI","type":"embedding","dimensions":1536,"maxInputTokens":8192,"pricePer1MTokens":0.02,"pricingNote":"$0.02 per 1M input tokens. The default budget choice for English RAG.","openSource":false,"license":"Proprietary","released":"2024-01-25","notes":"OpenAI budget embedding. 5x cheaper than ada-002 with better quality. Good default for most RAG agents.","multilingual":true,"url":"https://developers.openai.com/api/docs/guides/embeddings","mtebAvg":62.3}
+{"id":"voyage-4-large","name":"voyage-4-large","provider":"Voyage AI","type":"embedding","dimensions":1024,"maxInputTokens":32000,"pricePer1MTokens":0.12,"pricingNote":"$0.12 per 1M tokens after 200M free tokens per account. Batch API is 33% cheaper. Output dimensions 256, 512, 1024 (default), or 2048.","openSource":false,"license":"Proprietary","released":"2026-01-15","notes":"Voyage flagship on a mixture-of-experts architecture. Every voyage-4 model writes into the same embedding space, so documents indexed with voyage-4-large can be queried with the cheaper voyage-4-lite. A third cheaper than voyage-3-large.","multilingual":true,"url":"https://docs.voyageai.com/docs/embeddings","mtebAvg":null}
+{"id":"voyage-4","name":"voyage-4","provider":"Voyage AI","type":"embedding","dimensions":1024,"maxInputTokens":32000,"pricePer1MTokens":0.06,"pricingNote":"$0.06 per 1M tokens after 200M free tokens per account.","openSource":false,"license":"Proprietary","released":"2026-01-15","notes":"Mid-tier voyage-4 model and the current Voyage workhorse. Shares an embedding space with the rest of the voyage-4 series.","multilingual":true,"url":"https://docs.voyageai.com/docs/embeddings","mtebAvg":null}
+{"id":"voyage-4-lite","name":"voyage-4-lite","provider":"Voyage AI","type":"embedding","dimensions":1024,"maxInputTokens":32000,"pricePer1MTokens":0.02,"pricingNote":"$0.02 per 1M tokens after 200M free tokens per account. Up to 1M tokens per batch request.","openSource":false,"license":"Proprietary","released":"2026-01-15","notes":"Latency and cost tier of the voyage-4 series. The shared embedding space makes it a natural query-side model against a corpus indexed with voyage-4-large.","multilingual":true,"url":"https://docs.voyageai.com/docs/embeddings","mtebAvg":null}
+{"id":"voyage-code-4","name":"voyage-code-4","provider":"Voyage AI","type":"embedding","dimensions":1024,"maxInputTokens":32000,"pricePer1MTokens":0.12,"pricingNote":"$0.12 per 1M tokens, a third below voyage-code-3. Output dimensions 256, 512, 1024 (default), or 2048.","openSource":false,"license":"Proprietary","released":"2026-08-13","notes":"Code retrieval model tuned for coding-agent workloads. Voyage reports a 27.54% gain over voyage-code-3 on agentic retrieval and double-digit leads over Cohere Embed 4 and Gemini Embedding 2 on code search.","multilingual":true,"url":"https://docs.voyageai.com/docs/embeddings","mtebAvg":null}
+{"id":"voyage-4-nano","name":"voyage-4-nano","provider":"Voyage AI","type":"embedding","dimensions":2048,"maxInputTokens":32000,"pricePer1MTokens":null,"pricingNote":"Open weights on Hugging Face. Free to self-host.","openSource":true,"license":"Apache-2.0","released":"2026-01-15","notes":"The open-weight member of the voyage-4 series: 180M non-embedding plus 160M embedding parameters. Voyage says all voyage-4 embeddings are compatible with each other, so a self-hosted nano index can later be queried with the hosted voyage-4 models.","multilingual":true,"url":"https://huggingface.co/voyageai/voyage-4-nano","mtebAvg":null}
+{"id":"voyage-3-large","name":"voyage-3-large","provider":"Voyage AI","type":"embedding","dimensions":1024,"maxInputTokens":32000,"pricePer1MTokens":0.18,"pricingNote":"$0.18 per 1M input tokens. Listed under older models, with no free-token allowance.","openSource":false,"license":"Proprietary","released":"2025-01-07","notes":"Previous-generation Voyage flagship. voyage-4-large replaces it at $0.12, so new projects have no price reason to start here. Supports Matryoshka truncation.","multilingual":true,"url":"https://docs.voyageai.com/docs/embeddings","mtebAvg":67}
+{"id":"voyage-3","name":"voyage-3","provider":"Voyage AI","type":"embedding","dimensions":1024,"maxInputTokens":32000,"pricePer1MTokens":0.06,"pricingNote":"$0.06 per 1M input tokens. Listed under older models.","openSource":false,"license":"Proprietary","released":"2024-09-18","notes":"Older Voyage workhorse, superseded by voyage-3.5 and then voyage-4 at the same price. Keep only for existing indexes.","multilingual":true,"url":"https://docs.voyageai.com/docs/embeddings","mtebAvg":63.5}
+{"id":"voyage-3-lite","name":"voyage-3-lite","provider":"Voyage AI","type":"embedding","dimensions":512,"maxInputTokens":32000,"pricePer1MTokens":0.02,"pricingNote":"$0.02 per 1M input tokens. Listed under older models.","openSource":false,"license":"Proprietary","released":"2024-09-18","notes":"Older budget tier. 512-dim vectors keep storage costs low; voyage-4-lite is the current option at the same price.","multilingual":true,"url":"https://docs.voyageai.com/docs/embeddings","mtebAvg":60}
+{"id":"voyage-code-3","name":"voyage-code-3","provider":"Voyage AI","type":"embedding","dimensions":1024,"maxInputTokens":32000,"pricePer1MTokens":0.18,"pricingNote":"$0.18 per 1M input tokens. Listed under older models.","openSource":false,"license":"Proprietary","released":"2024-12-05","notes":"Previous code-search model, replaced by voyage-code-4 at a lower price.","multilingual":true,"url":"https://docs.voyageai.com/docs/embeddings","mtebAvg":null}
+{"id":"embed-v4.0","name":"embed-v4.0","provider":"Cohere","type":"embedding","dimensions":1536,"maxInputTokens":128000,"pricePer1MTokens":0.12,"pricingNote":"$0.12 per 1M text tokens, $0.47 per 1M image tokens. Output dimensions 256, 512, 1024, or 1536 (default).","openSource":false,"license":"Proprietary","released":"2025-04-15","notes":"Cohere flagship embedding and the only embedding model on its public price list. Embeds text, images, and mixed documents such as PDFs into one space, 100+ languages, and a 128K context that removes most chunking pressure.","multilingual":true,"url":"https://docs.cohere.com/docs/cohere-embed","mtebAvg":null}
+{"id":"embed-multilingual-v3.0","name":"embed-multilingual-v3.0","provider":"Cohere","type":"embedding","dimensions":1024,"maxInputTokens":512,"pricePer1MTokens":0.1,"pricingNote":"$0.10 per 1M input tokens at last listing. Cohere's public pricing page now shows only Embed 4, so confirm the v3 rate in the dashboard.","openSource":false,"license":"Proprietary","released":"2023-11-02","notes":"Previous Cohere multilingual model, still served. 100+ languages with strong cross-lingual retrieval, but the 512-token input limit is the main constraint next to embed-v4.0.","multilingual":true,"url":"https://docs.cohere.com/docs/cohere-embed","mtebAvg":64}
+{"id":"embed-english-v3.0","name":"embed-english-v3.0","provider":"Cohere","type":"embedding","dimensions":1024,"maxInputTokens":512,"pricePer1MTokens":0.1,"pricingNote":"$0.10 per 1M input tokens at last listing. Not shown on the current public pricing page.","openSource":false,"license":"Proprietary","released":"2023-11-02","notes":"English-only sibling of embed-multilingual-v3. Slightly stronger on English-only corpora.","multilingual":false,"url":"https://docs.cohere.com/docs/cohere-embed","mtebAvg":64.5}
+{"id":"gemini-embedding-2","name":"gemini-embedding-2","provider":"Google","type":"embedding","dimensions":3072,"maxInputTokens":8192,"pricePer1MTokens":0.2,"pricingNote":"$0.20 per 1M text tokens on the Gemini API and Vertex AI; images $0.45, audio $6.50, video $12.00 per 1M tokens. Batch is half price on Vertex AI.","openSource":false,"license":"Proprietary","released":"2026-04-22","notes":"Google's first multimodal embedding model, mapping text, images, video, audio, and PDFs into one space. Matryoshka output from 128 to 3072 dimensions, and unlike 001 it re-normalizes truncated vectors automatically. Replaces gemini-embedding-001 as Google's recommended model.","multilingual":true,"url":"https://ai.google.dev/gemini-api/docs/embeddings","mtebAvg":null}
+{"id":"gemini-embedding-001","name":"gemini-embedding-001","provider":"Google","type":"embedding","dimensions":3072,"maxInputTokens":2048,"pricePer1MTokens":0.15,"pricingNote":"$0.15 per 1M input tokens on Vertex AI ($0.12 batch). The Gemini API pricing page now lists only Gemini Embedding 2. Supports Matryoshka.","openSource":false,"license":"Proprietary","released":"2025-07-14","notes":"Previous Google text embedding, still served with a long runway: earliest shutdown 2028-05-14 on the Gemini API and no sooner than 2028-05-20 on Vertex AI. Strong multilingual. Truncated outputs need manual re-normalization.","multilingual":true,"url":"https://ai.google.dev/gemini-api/docs/embeddings","mtebAvg":68.3}
+{"id":"text-embedding-005","name":"text-embedding-005","provider":"Google","type":"embedding","dimensions":768,"maxInputTokens":2048,"pricePer1MTokens":0.025,"pricingNote":"$0.000025 per 1,000 input units on Vertex AI ($0.00002 batch). Vertex AI only.","openSource":false,"license":"Proprietary","released":"2024-11-18","notes":"Google budget tier. 768-dim, English-focused, cheapest of the Vertex AI embeddings. Scheduled to retire on Vertex AI on 2027-04-01, so avoid it for new long-lived indexes.","multilingual":false,"url":"https://cloud.google.com/vertex-ai/generative-ai/docs/embeddings","mtebAvg":null}
+{"id":"embeddinggemma-300m","name":"embeddinggemma-300m","provider":"Google","type":"embedding","dimensions":768,"maxInputTokens":2048,"pricePer1MTokens":null,"pricingNote":"Open weights on Hugging Face (license acceptance required). Free to self-host.","openSource":true,"license":"Gemma Terms of Use","released":"2025-09-04","notes":"Google's 308M-parameter open embedding model for on-device and offline RAG. Matryoshka truncation to 512, 256, or 128 dimensions. MTEB English v2 69.67 and Multilingual v2 61.15 at 768 dimensions (v2 scores).","multilingual":true,"url":"https://ai.google.dev/gemma/docs/embeddinggemma","mtebAvg":69.67}
+{"id":"mistral-embed","name":"mistral-embed","provider":"Mistral","type":"embedding","dimensions":1024,"maxInputTokens":8000,"pricePer1MTokens":0.1,"pricingNote":"$0.10 per 1M input tokens via la Plateforme.","openSource":false,"license":"Proprietary","released":"2023-12-11","notes":"European data residency option. Solid English/French/German performance. Same pricing tier as Cohere but with longer context. For code, Mistral sells Codestral Embed at $0.15 per 1M tokens.","multilingual":true,"url":"https://docs.mistral.ai/models","mtebAvg":60.7}
+{"id":"jina-embeddings-v5-text-small","name":"jina-embeddings-v5-text-small","provider":"Jina AI","type":"embedding","dimensions":1024,"maxInputTokens":32768,"pricePer1MTokens":0.05,"pricingNote":"$0.05 per 1M tokens via the Jina API after a one-time 10M free tokens. Also served through Elastic Inference Service.","openSource":true,"license":"CC-BY-NC-4.0","released":"2026-02-18","notes":"677M multilingual model built on Qwen3-0.6B and distilled from Qwen3-Embedding-4B, covering 119+ languages with Matryoshka dimensions from 32 to 1024. MTEB English v2 71.7 and MMTEB 67.7 (v2 scores). Weights are non-commercial; commercial self-hosting goes through Elastic, which now sells Jina models.","multilingual":true,"url":"https://huggingface.co/jinaai/jina-embeddings-v5-text-small","mtebAvg":71.7}
+{"id":"jina-embeddings-v3","name":"jina-embeddings-v3","provider":"Jina AI","type":"embedding","dimensions":1024,"maxInputTokens":8192,"pricePer1MTokens":0.05,"pricingNote":"$0.05 per 1M tokens via the Jina API. Weights are CC-BY-NC-4.0, so commercial self-hosting needs a license from Elastic.","openSource":true,"license":"CC-BY-NC-4.0","released":"2024-09-18","notes":"Previous-generation open-weights multilingual embedding, superseded by jina-embeddings-v5. Adapter-based: the same model serves retrieval, classification, and separation tasks via task-specific LoRAs.","multilingual":true,"url":"https://jina.ai/embeddings/","mtebAvg":65.5}
+{"id":"qwen3-embedding-8b","name":"Qwen3-Embedding-8B","provider":"Alibaba Qwen","type":"embedding","dimensions":4096,"maxInputTokens":32000,"pricePer1MTokens":null,"pricingNote":"Open weights. Free to self-host. 0.6B and 4B siblings share the recipe.","openSource":true,"license":"Apache-2.0","released":"2025-06-05","notes":"The open-weight embedding most teams benchmark against. Instruction-aware, 100+ natural and programming languages, output dimensions from 32 to 4096. MTEB English v2 75.22 and MTEB Multilingual 70.58 (v2 scores). The 0.6B version is the most downloaded model in the family for cheap self-hosting.","multilingual":true,"url":"https://huggingface.co/Qwen/Qwen3-Embedding-8B","mtebAvg":75.22}
+{"id":"nemotron-3-embed-1b","name":"Nemotron-3-Embed-1B","provider":"NVIDIA","type":"embedding","dimensions":2048,"maxInputTokens":32768,"pricePer1MTokens":null,"pricingNote":"Open weights on Hugging Face. Free to self-host.","openSource":true,"license":"OpenMDW-1.1","released":"2026-07-16","notes":"1.14B multilingual retrieval embedding evaluated across 34 languages, with slicing to 1024 or 512 dimensions after re-normalization. NVIDIA reports RTEB 72.38 and MMTEB retrieval 71.04 NDCG@10, against 61.98 and 59.71 for llama-nemotron-embed-vl-1b-v2 on the same tables. Retrieval-focused, so there is no full MTEB average.","multilingual":true,"url":"https://huggingface.co/nvidia/Nemotron-3-Embed-1B-BF16","mtebAvg":null}
+{"id":"nomic-embed-text-v1.5","name":"nomic-embed-text-v1.5","provider":"Nomic AI","type":"embedding","dimensions":768,"maxInputTokens":8192,"pricePer1MTokens":null,"pricingNote":"Open weights. Free to self-host. Hosted through Nomic Atlas; Nomic does not publish a per-token rate.","openSource":true,"license":"Apache-2.0","released":"2024-02-14","notes":"Open-source English embedding with Matryoshka support (truncate to 256/512/768). Reproducible training data; one of the few fully open embeddings.","multilingual":false,"url":"https://blog.nomic.ai/posts/nomic-embed-matryoshka","mtebAvg":62.3}
+{"id":"mxbai-embed-large-v1","name":"mxbai-embed-large-v1","provider":"Mixedbread","type":"embedding","dimensions":1024,"maxInputTokens":512,"pricePer1MTokens":null,"pricingNote":"Open weights. Free to self-host. Also available through the paid Mixedbread API.","openSource":true,"license":"Apache-2.0","released":"2024-03-07","notes":"Apache-licensed dense retriever. Strong English performance. Short input limit (512) is the main constraint vs Voyage/Jina.","multilingual":false,"url":"https://www.mixedbread.com/docs/embeddings/overview","mtebAvg":64.7}
+{"id":"bge-m3","name":"bge-m3","provider":"BAAI","type":"embedding","dimensions":1024,"maxInputTokens":8192,"pricePer1MTokens":null,"pricingNote":"Open weights. Free to self-host. Available through most inference providers (Together, DeepInfra, Replicate).","openSource":true,"license":"MIT","released":"2024-01-30","notes":"Multi-functional, multi-lingual, multi-granularity. Outputs dense + sparse + multi-vector representations from a single model. Strong on long-document multilingual retrieval.","multilingual":true,"url":"https://huggingface.co/BAAI/bge-m3","mtebAvg":64.5}
+{"id":"rerank-v4.0-pro","name":"rerank-v4.0-pro","provider":"Cohere","type":"reranker","dimensions":null,"maxInputTokens":32768,"pricePer1MTokens":0,"pricingNote":"$2.50 per 1k searches.","openSource":false,"license":"Proprietary","released":"2025-12-11","notes":"Cohere flagship reranker with a 32K context, four times rerank-v3.5, and support for semi-structured JSON documents. The quality tier for complex enterprise retrieval.","multilingual":true,"url":"https://docs.cohere.com/docs/rerank","mtebAvg":null}
+{"id":"rerank-v4.0-fast","name":"rerank-v4.0-fast","provider":"Cohere","type":"reranker","dimensions":null,"maxInputTokens":32768,"pricePer1MTokens":0,"pricingNote":"$2 per 1k searches.","openSource":false,"license":"Proprietary","released":"2025-12-11","notes":"Lighter Rerank 4 variant for latency-sensitive and high-throughput pipelines, with the same 32K context as Pro.","multilingual":true,"url":"https://docs.cohere.com/docs/rerank","mtebAvg":null}
+{"id":"rerank-v3.5","name":"rerank-v3.5","provider":"Cohere","type":"reranker","dimensions":null,"maxInputTokens":4096,"pricePer1MTokens":0,"pricingNote":"$2 per 1k searches at last listing. Cohere's public pricing page now lists only Rerank 4 per-search rates.","openSource":false,"license":"Proprietary","released":"2024-12-03","notes":"Previous Cohere reranker, still served on the Cohere API. Rerank 4 quadruples the context window. Azure retired its hosted copy on 2026-05-14.","multilingual":true,"url":"https://docs.cohere.com/docs/rerank","mtebAvg":null}
+{"id":"rerank-2.5","name":"rerank-2.5","provider":"Voyage AI","type":"reranker","dimensions":null,"maxInputTokens":32000,"pricePer1MTokens":0.05,"pricingNote":"$0.05 per 1M tokens (query tokens times documents plus document tokens).","openSource":false,"license":"Proprietary","released":"2025-08-11","notes":"Voyage's recommended reranker, with instruction-following and a 32K query-plus-document limit. rerank-2.5-lite is $0.02 per 1M tokens. A rerank-3 preview is also listed at the same price.","multilingual":true,"url":"https://docs.voyageai.com/docs/reranker","mtebAvg":null}
+{"id":"rerank-2","name":"rerank-2","provider":"Voyage AI","type":"reranker","dimensions":null,"maxInputTokens":16000,"pricePer1MTokens":0.05,"pricingNote":"$0.05 per 1M tokens (query + documents combined). Listed under older models.","openSource":false,"license":"Proprietary","released":"2024-08-09","notes":"Previous Voyage reranker, superseded by rerank-2.5 at the same price with double the context.","multilingual":true,"url":"https://docs.voyageai.com/docs/reranker","mtebAvg":null}
+{"id":"jina-reranker-v3.5","name":"jina-reranker-v3.5","provider":"Jina AI","type":"reranker","dimensions":null,"maxInputTokens":131072,"pricePer1MTokens":0.05,"pricingNote":"$0.05 per 1M tokens via the Jina API.","openSource":true,"license":"CC-BY-NC-4.0","released":"2026-07-27","notes":"Listwise reranker on Qwen3-0.6B that scores a whole candidate list jointly within a 131K context. Jina reports it beats a 4B Qwen reranker on BEIR at about a seventh of the size. Non-commercial weights.","multilingual":true,"url":"https://huggingface.co/jinaai/jina-reranker-v3.5","mtebAvg":null}
+{"id":"jina-reranker-v2","name":"jina-reranker-v2-base-multilingual","provider":"Jina AI","type":"reranker","dimensions":null,"maxInputTokens":1024,"pricePer1MTokens":0.05,"pricingNote":"$0.05 per 1M tokens via Jina API. Weights are CC-BY-NC-4.0.","openSource":true,"license":"CC-BY-NC-4.0","released":"2024-07-04","notes":"Older open-weights reranker, 100+ languages. Superseded by jina-reranker-v3 and v3.5 at the same API price with far longer context.","multilingual":true,"url":"https://jina.ai/reranker/","mtebAvg":null}
+{"id":"qwen3-reranker-4b","name":"Qwen3-Reranker-4B","provider":"Alibaba Qwen","type":"reranker","dimensions":null,"maxInputTokens":32000,"pricePer1MTokens":null,"pricingNote":"Open weights. Free to self-host. 0.6B and 8B sizes also available.","openSource":true,"license":"Apache-2.0","released":"2025-06-05","notes":"The most downloaded open reranker on Hugging Face and the commercially clean alternative to the non-commercial Jina weights. Instruction-aware, 100+ languages, 32K context. Pairs with Qwen3-Embedding.","multilingual":true,"url":"https://huggingface.co/Qwen/Qwen3-Reranker-4B","mtebAvg":null}

2026-10-12/embodied-ai.jsonl ADDED Viewed

	@@ -0,0 +1,34 @@

+{"id":"pi-0","name":"pi-0","org":"Physical Intelligence","category":"foundation_model","parameters":"3.3B","released":"2024-10","license":"Apache-2.0 (open weights)","paperUrl":"https://www.pi.website/blog/pi0","codeUrl":"https://github.com/Physical-Intelligence/openpi","demoUrl":"https://www.pi.website","notes":"Generalist VLA built on PaliGemma. Trained on 10K+ hours of cross-embodiment teleoperation data. The open release that broadly shifted the field toward VLA foundation models."}
+{"id":"pi-0.5","name":"pi-0.5","org":"Physical Intelligence","category":"foundation_model","parameters":"3.3B","released":"2025-04","license":"Apache-2.0 (open weights)","paperUrl":"https://www.pi.website/blog/pi05","codeUrl":"https://github.com/Physical-Intelligence/openpi","demoUrl":"https://www.pi.website","notes":"Successor to pi-0 with open-world generalization to homes the model never saw during training. Base checkpoint added to the openpi repo in September 2025."}
+{"id":"pi-0.7","name":"pi-0.7","org":"Physical Intelligence","category":"foundation_model","parameters":"5B","released":"2026-04","license":"proprietary (weights not released)","paperUrl":"https://www.pi.website/blog/pi07","codeUrl":null,"demoUrl":"https://www.pi.website","notes":"Steerable generalist VLA on a Gemma 3 4B backbone with an 860M flow-matching action expert. Conditions on strategy metadata and subgoal images, and matched Physical Intelligence's own RL-finetuned specialists on laundry, espresso and box building."}
+{"id":"gr00t-n1","name":"GR00T N1","org":"NVIDIA","category":"foundation_model","parameters":"2B","released":"2025-03","license":"open weights","paperUrl":"https://arxiv.org/abs/2503.14734","codeUrl":"https://huggingface.co/nvidia/GR00T-N1-2B","demoUrl":"https://developer.nvidia.com/isaac/gr00t","notes":"NVIDIA dual-system VLA targeting humanoids: System-2 reasoning over a vision-language backbone, System-1 diffusion policy. First open NVIDIA humanoid foundation model."}
+{"id":"gr00t-n1.7","name":"GR00T N1.7","org":"NVIDIA","category":"foundation_model","parameters":"3B","released":"2026-04","license":"open weights, commercial use permitted (NVIDIA Open Model License; code Apache-2.0)","paperUrl":"https://huggingface.co/blog/nvidia/gr00t-n1-7","codeUrl":"https://huggingface.co/nvidia/GR00T-N1.7-3B","demoUrl":"https://developer.nvidia.com/isaac/gr00t","notes":"Swaps the Eagle backbone for Cosmos-Reason2-2B (Qwen3-VL architecture) and pretrains on 20K hours of EgoScale human video using a relative end-effector action space. First GR00T release licensed for commercial deployment; shipped as early access in April 2026 before reaching general availability."}
+{"id":"rt-2","name":"RT-2","org":"Google DeepMind","category":"foundation_model","parameters":"55B","released":"2023-07","license":"proprietary","paperUrl":"https://robotics-transformer2.github.io","codeUrl":null,"demoUrl":null,"notes":"First VLM-to-action transformer. PaLI-X (55B) and PaLM-E backbones output robot actions as tokens. Closed but seeded the entire VLA category."}
+{"id":"gemini-robotics-2","name":"Gemini Robotics 2","org":"Google DeepMind","category":"foundation_model","parameters":"undisclosed","released":"2026-07","license":"proprietary (early-access partners)","paperUrl":"https://deepmind.google/blog/gemini-robotics-2-brings-whole-body-intelligence-to-robots/","codeUrl":null,"demoUrl":"https://deepmind.google/models/gemini-robotics/on-device/","notes":"VLA that controls full humanoids from feet to fingertips, demonstrated on Apptronik Apollo 2, plus bi-arm robots from one checkpoint. Ships alongside Gemini Robotics On-Device 2, a local VLA that adapts to new bi-arm embodiments with typically fewer than 200 examples."}
+{"id":"gemini-robotics-er-2","name":"Gemini Robotics ER 2","org":"Google DeepMind","category":"foundation_model","parameters":"undisclosed","released":"2026-07","license":"proprietary (Gemini API, public preview)","paperUrl":"https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-robotics-er-2/","codeUrl":null,"demoUrl":"https://deepmind.google/models/gemini-robotics/embodied-reasoning/","notes":"Embodied reasoning VLM built on Gemini 3.5 Flash that acts as a robot's high-level brain: plans multi-step tasks, tracks progress from video, calls tools, and hands motor execution to any VLA. Adds multi-robot collaboration and is available to developers via the Gemini API and Google AI Studio."}
+{"id":"openvla","name":"OpenVLA","org":"Stanford / Berkeley / Toyota Research","category":"foundation_model","parameters":"7B","released":"2024-06","license":"MIT","paperUrl":"https://openvla.github.io","codeUrl":"https://github.com/openvla/openvla","demoUrl":"https://huggingface.co/openvla/openvla-7b","notes":"Open replication of RT-2 trained on Open X-Embodiment. The reference open VLA before pi-0; still widely fine-tuned for academic robotics work."}
+{"id":"octo","name":"Octo","org":"Berkeley AI Research","category":"foundation_model","parameters":"93M","released":"2024-05","license":"MIT","paperUrl":"https://octo-models.github.io","codeUrl":"https://github.com/octo-models/octo","demoUrl":null,"notes":"Small generalist transformer policy (93M and 27M variants) trained on 800K Open X-Embodiment trajectories. Designed as a fine-tunable base."}
+{"id":"rdt-1b","name":"RDT-1B","org":"Tsinghua University","category":"foundation_model","parameters":"1.2B","released":"2024-10","license":"MIT","paperUrl":"https://rdt-robotics.github.io/rdt-robotics/","codeUrl":"https://github.com/thu-ml/RoboticsDiffusionTransformer","demoUrl":null,"notes":"Largest open bimanual manipulation diffusion transformer at release. Pretrained on 1M+ multi-robot episodes, fine-tuned for bimanual ALOHA-class hardware."}
+{"id":"helix","name":"Helix","org":"Figure AI","category":"foundation_model","parameters":"undisclosed","released":"2025-02","license":"proprietary","paperUrl":"https://www.figure.ai/news/helix","codeUrl":null,"demoUrl":"https://www.figure.ai","notes":"Dual-system VLA launched on Figure 02 with high-rate control of the full upper body, including wrists, torso, head and individual fingers. Runs entirely onboard embedded low-power GPUs; succeeded by Helix 02 in January 2026."}
+{"id":"helix-02","name":"Helix 02","org":"Figure AI","category":"foundation_model","parameters":"undisclosed","released":"2026-01","license":"proprietary","paperUrl":"https://www.figure.ai/news/helix-02","codeUrl":null,"demoUrl":"https://www.figure.ai","notes":"Extends Helix to full-body loco-manipulation from pixels by adding System 0, a 10M-parameter whole-body controller running at 1 kHz trained on 1,000+ hours of human motion data. Demonstrated a four-minute autonomous dishwasher unload and reload on Figure 03."}
+{"id":"skild-s1","name":"S1","org":"Skild AI","category":"foundation_model","parameters":"undisclosed","released":"2026-08","license":"proprietary","paperUrl":"https://www.skild.ai/blogs/s1","codeUrl":null,"demoUrl":"https://www.skild.ai","notes":"Robotic foundation model built as an in-context learner: a single video demonstration specifies the task, with no fine-tuning or post-training. Skild reports executing unseen manipulation tasks up to 10 minutes long."}
+{"id":"figure-02","name":"Figure 02","org":"Figure AI","category":"humanoid","parameters":null,"released":"2024-08","license":"commercial","paperUrl":"https://www.prnewswire.com/news-releases/figure-unveils-figure-02-its-second-generation-humanoid-setting-new-standards-in-ai-and-robotics-302214889.html","codeUrl":null,"demoUrl":"https://www.figure.ai","notes":"Second-generation Figure humanoid with 16-DoF hands and 3x the onboard compute of Figure 01. Ran an 11-month deployment at BMW Plant Spartanburg, contributing to 30,000+ X3 vehicles, before Figure began retiring it in November 2025 in favor of Figure 03."}
+{"id":"figure-03","name":"Figure 03","org":"Figure AI","category":"humanoid","parameters":null,"released":"2025-10","license":"commercial","paperUrl":"https://www.figure.ai/news/introducing-figure-03","codeUrl":null,"demoUrl":"https://www.figure.ai/figure","notes":"Third-generation humanoid redesigned for Helix, the home and volume manufacturing at BotQ: palm cameras, fingertip tactile sensors that register about 3 grams, soft washable coverings. Arrived at BMW Plant Spartanburg in June 2026 running Helix 02."}
+{"id":"1x-neo","name":"1X NEO","org":"1X Technologies","category":"humanoid","parameters":null,"released":"2025-10","license":"commercial (consumer)","paperUrl":"https://www.1x.tech/discover/neo-home-robot","codeUrl":null,"demoUrl":"https://www.1x.tech","notes":"Home humanoid that 1X bills as the first consumer-ready humanoid: 30 kg (66 lb), tendon-driven joints, 22-DoF hands, $20K early access or $499/month. The Hayward factory began full-scale production in April 2026, with first US deliveries promised for 2026."}
+{"id":"tesla-optimus-gen3","name":"Optimus","org":"Tesla","category":"humanoid","parameters":null,"released":"2023-12","license":"commercial","paperUrl":"https://assets-ir.tesla.com/tesla-contents/IR/TSLA-Q2-2026-Update.pdf","codeUrl":null,"demoUrl":"https://www.tesla.com/AI","notes":"Gen 2 (December 2023) remains the latest body Tesla has publicly unveiled; the V3 reveal slipped from Q1 2026 and had not happened by September 2026. Tesla's Q2 2026 update says first-generation Optimus lines are being installed at Fremont, with initial builds going to its Optimus Academy for training data collection."}
+{"id":"apptronik-apollo","name":"Apollo 2","org":"Apptronik","category":"humanoid","parameters":null,"released":"2026-06","license":"commercial","paperUrl":"https://apptronik.com/news-collection/welcome-to-robot-park-where-apptroniks-apollo-goes-to-work","codeUrl":null,"demoUrl":"https://apptronik.com/apollo/apollo-2","notes":"Modular humanoid in bipedal and wheeled-base configurations, unveiled with Apptronik's Robot Park data-collection facility in Austin. Its fleet data trains Google DeepMind's Gemini Robotics models; Mercedes-Benz and GXO are customers, and a commercial Apollo 3 is in development."}
+{"id":"unitree-g1","name":"Unitree G1","org":"Unitree","category":"humanoid","parameters":null,"released":"2024-05","license":"commercial","paperUrl":"https://www.unitree.com/g1","codeUrl":"https://github.com/unitreerobotics","demoUrl":"https://www.unitree.com/g1","notes":"Compact 35 kg research humanoid (23 to 43 DoF) launched at $16K and now listed from $13.5K. Popular for academic research; NVIDIA's Isaac GR00T platform supports it. SDK and Python bindings published."}
+{"id":"unitree-h1","name":"Unitree H1","org":"Unitree","category":"humanoid","parameters":null,"released":"2023-08","license":"commercial","paperUrl":"https://www.unitree.com/h1","codeUrl":"https://github.com/unitreerobotics","demoUrl":"https://www.unitree.com/h1","notes":"Unitree's first full-size humanoid (about 180 cm, 47 kg), rated at 3.3 m/s, which Unitree lists as a world record. Common research platform for whole-body control papers; the 31-DoF H2 now sits above it in the lineup."}
+{"id":"boston-dynamics-atlas-electric","name":"Atlas (Electric)","org":"Boston Dynamics","category":"humanoid","parameters":null,"released":"2026-01","license":"commercial (enterprise)","paperUrl":"https://bostondynamics.com/blog/boston-dynamics-unveils-new-atlas-robot-to-revolutionize-industry/","codeUrl":null,"demoUrl":"https://bostondynamics.com/products/atlas/","notes":"Product version of the all-electric Atlas, unveiled at CES 2026: 56 DoF, 1.9 m, 90 kg, 50 kg instant lift, self-swapping batteries, IP67. All 2026 units are committed to Hyundai's RMAC and Google DeepMind, which is integrating Gemini Robotics models."}
+{"id":"agility-digit","name":"Digit","org":"Agility Robotics","category":"humanoid","parameters":null,"released":"2023-03","license":"commercial","paperUrl":"https://www.agilityrobotics.com/content/agility-robotics-launches-next-generation-of-digit-worlds-first-human-centric-multi-purpose-robot-made-for-logistics-work","codeUrl":null,"demoUrl":"https://www.agilityrobotics.com","notes":"Bipedal humanoid built for warehouse tote handling, sold via RaaS or direct purchase, with early trials at Amazon and GXO and deployments at Schaeffler, Toyota and Mercado Libre. Agility plans first customer deliveries of Digit v5 (50 lb payload, cooperative safety) in December 2026."}
+{"id":"sanctuary-phoenix","name":"Phoenix","org":"Sanctuary AI","category":"humanoid","parameters":null,"released":"2023-05","license":"commercial","paperUrl":"https://sanctuary.ai/news/sanctuary-ai-unveils-phoenix-a-humanoid-general-purpose-robot-designed-for-work/","codeUrl":null,"demoUrl":"https://sanctuary.ai","notes":"Sixth-generation Sanctuary general-purpose robot running its Carbon AI control system, launched with 20-DoF hands and proprietary haptic sensing. Sanctuary now focuses on deploying its Physical AI and hydraulic hands across commercially available robot hardware."}
+{"id":"open-x-embodiment","name":"Open X-Embodiment","org":"Google DeepMind + 21 institutions","category":"dataset","parameters":null,"released":"2023-10","license":"CC-BY-4.0","paperUrl":"https://robotics-transformer-x.github.io","codeUrl":"https://github.com/google-deepmind/open_x_embodiment","demoUrl":"https://huggingface.co/datasets/jxu124/OpenX-Embodiment","notes":"Standardized cross-embodiment dataset spanning 22 robot embodiments and 1M+ trajectories. The pretraining substrate behind Octo, OpenVLA, and many followups."}
+{"id":"droid","name":"DROID","org":"Stanford / Berkeley / Toyota Research / Google","category":"dataset","parameters":null,"released":"2024-03","license":"CC-BY-4.0","paperUrl":"https://droid-dataset.github.io","codeUrl":"https://github.com/droid-dataset/droid","demoUrl":"https://huggingface.co/datasets/KarlP/droid","notes":"76K teleoperated trajectories (about 350 hours) across 564 scenes and 86 tasks, all on Franka arms, gathered by 50 collectors on three continents. Part of the open-source slice of the pi-0 pretraining mixture."}
+{"id":"agibot-world","name":"AgiBot World","org":"AgiBot","category":"dataset","parameters":null,"released":"2025-03","license":"CC-BY-NC-SA-4.0","paperUrl":"https://agibot-world.com","codeUrl":"https://huggingface.co/datasets/agibot-world/AgiBotWorld-Beta","demoUrl":"https://agibot-world.com","notes":"Beta release holds 1M+ trajectories (2,976 hours) from 100 robots across 100+ real-world scenarios, covering contact-rich manipulation, long-horizon planning and multi-robot collaboration. Followed the December 2024 Alpha release."}
+{"id":"mobile-aloha","name":"Mobile ALOHA","org":"Stanford","category":"dataset","parameters":null,"released":"2024-01","license":"MIT","paperUrl":"https://mobile-aloha.github.io","codeUrl":"https://github.com/MarkFzp/mobile-aloha","demoUrl":"https://mobile-aloha.github.io","notes":"Open hardware bimanual mobile manipulator built on a $32K budget, plus teleoperation data and training code. With 50 demonstrations per task, co-training with static ALOHA data raised success rates by up to 90%."}
+{"id":"bridgedata-v2","name":"BridgeData V2","org":"Berkeley AI Research","category":"dataset","parameters":null,"released":"2023-08","license":"CC-BY-4.0","paperUrl":"https://rail-berkeley.github.io/bridgedata","codeUrl":"https://github.com/rail-berkeley/bridge_data_v2","demoUrl":null,"notes":"60K+ trajectories on WidowX arms across 24 environments. Foundational for skill generalization research; included in Open X-Embodiment."}
+{"id":"hiw-500","name":"HIW-500 (Humanoids In-the-Wild)","org":"BitRobot / Unitree / Hugging Face","category":"dataset","parameters":null,"released":"2026-06","license":"CC-BY-4.0","paperUrl":"https://bitrobot-foundation.github.io/humanoids-in-the-wild-500-hours/","codeUrl":"https://huggingface.co/datasets/BitRobot/HIW-500-LeRobot","demoUrl":"https://huggingface.co/datasets/BitRobot/HIW-500","notes":"500+ hours and 23K+ episodes of whole-body Unitree G1 teleoperation in 12 real homes in Southeast Asia, with 148K+ subtask annotations. Published as raw ROS bag / MCAP recordings plus a LeRobot-format conversion."}
+{"id":"isaac-lab","name":"Isaac Lab","org":"NVIDIA","category":"simulator","parameters":null,"released":"2025-01","license":"BSD-3-Clause","paperUrl":"https://isaac-sim.github.io/IsaacLab/","codeUrl":"https://github.com/isaac-sim/IsaacLab","demoUrl":"https://developer.nvidia.com/isaac/sim","notes":"GPU-accelerated robot learning framework on Isaac Sim. Supersedes Isaac Gym. Used for sim-to-real RL training of locomotion and manipulation policies; Isaac Lab 2.0 shipped January 2025 and a 3.0 beta integrating Newton followed in March 2026."}
+{"id":"newton","name":"Newton","org":"Linux Foundation (NVIDIA / Google DeepMind / Disney Research)","category":"simulator","parameters":null,"released":"2026-03","license":"Apache-2.0","paperUrl":"https://developer.nvidia.com/blog/newton-adds-contact-rich-manipulation-and-locomotion-capabilities-for-industrial-robotics/","codeUrl":"https://github.com/newton-physics/newton","demoUrl":"https://newton-physics.github.io/newton/stable/","notes":"Open GPU physics engine built on NVIDIA Warp and OpenUSD with MuJoCo Warp as its primary solver, plus Disney's Kamino and deformable solvers. Reached 1.0 GA at GTC 2026 and now ships monthly releases."}
+{"id":"mujoco-playground","name":"MuJoCo Playground","org":"Google DeepMind","category":"simulator","parameters":null,"released":"2025-01","license":"Apache-2.0","paperUrl":"https://playground.mujoco.org","codeUrl":"https://github.com/google-deepmind/mujoco_playground","demoUrl":"https://playground.mujoco.org","notes":"JAX-based MuJoCo training suite with battery-included locomotion and manipulation environments. Designed for massively parallel sim-to-real RL."}
+{"id":"genesis","name":"Genesis World","org":"Genesis AI","category":"simulator","parameters":null,"released":"2026-05","license":"Apache-2.0","paperUrl":"https://genesis.ai/blog/the-role-of-simulation-in-scalable-robotics-genesis-world-10-and-the-path-forward","codeUrl":"https://github.com/Genesis-Embodied-AI/genesis-world","demoUrl":"https://genesis-world.readthedocs.io","notes":"Unified multi-physics simulator (rigid, FEM, MPM, particle) that started as the academic Genesis project in December 2024. Renamed Genesis World at 1.0, adding the Nyx path-traced renderer and Quadrants Python-to-GPU compiler, with development now backed by Genesis AI."}

2026-10-12/fine-tuning.jsonl ADDED Viewed

	@@ -0,0 +1,13 @@

+{"id":"openai-finetune","name":"OpenAI Fine-Tuning","vendor":"OpenAI","type":"first-party","baseModels":["GPT-4.1","GPT-4.1-mini","GPT-4.1-nano (base retires 2026-10-23)","GPT-4o (2024-08-06, vision)","GPT-4o-mini","o4-mini (RFT only; base retires 2026-10-23)"],"methods":["full","dpo","rlhf"],"trainingPricing":"$25/1M tokens (GPT-4.1, GPT-4o); $5 (GPT-4.1-mini); $3 (GPT-4o-mini); $1.50 (GPT-4.1-nano); o4-mini reinforcement fine-tuning $100 per training hour","inferencePricing":"Fine-tuned GPT-4.1 $3.00 in / $12.00 out per 1M tokens; GPT-4.1-mini $0.80 / $3.20; o4-mini $4.00 / $16.00, half with data sharing","freeTier":null,"features":["Supervised, DPO, vision, and reinforcement fine-tuning","Closed to new organizations since 2026-05-07","Fine-tunes keep serving until their base model is deprecated"],"url":"https://developers.openai.com/api/docs/deprecations#update-to-openais-self-serve-fine-tuning","notes":"OpenAI is winding the platform down. New organizations lost access on 2026-05-07, organizations with no fine-tuned inference in the prior 60 days lost job creation on 2026-07-02, and every remaining customer stops creating jobs on 2027-01-06. Tunes built on gpt-4.1-nano and o4-mini stop serving when those bases shut down on 2026-10-23. No model export. Useful now only for maintaining existing tunes."}
+{"id":"anthropic-finetune","name":"Anthropic Fine-Tuning (AWS Bedrock)","vendor":"Anthropic","type":"first-party","baseModels":["Claude 3 Haiku (Bedrock end of life 2026-09-10)"],"methods":["lora"],"trainingPricing":"Not available for new jobs","inferencePricing":"Existing custom models ran on Provisioned Throughput, which cannot be newly created for Legacy models","freeTier":null,"features":["Bedrock-only (AWS)","Claude 3 Haiku was the only customizable Claude model","Legacy status blocks new fine-tuning jobs"],"url":"https://docs.aws.amazon.com/bedrock/latest/userguide/custom-model-fine-tuning.html","notes":"Effectively closed. Bedrock's fine-tuning table lists Claude 3 Haiku as the only customizable Claude model; it entered Legacy on 2026-03-10, which blocks new fine-tuning jobs, and reached end of life on 2026-09-10. No Claude 4 or Claude 5 model appears in the table."}
+{"id":"google-vertex-finetune","name":"Google Vertex AI Fine-Tuning","vendor":"Google","type":"first-party","baseModels":["Gemini 3.5 Flash","Gemini 3.1 Flash-Lite","Gemini 2.5 Pro, Flash, and Flash-Lite (retire 2026-10-20)","Gemma 3 (1B to 27B)","Llama 3.1, 3.2, 3.3, and Llama 4 Scout","Qwen3 4B to 32B"],"methods":["lora","dpo","rlhf"],"trainingPricing":"$10/1M training tokens (Gemini 3.5 Flash, SFT or RL); $3 (Gemini 3.1 Flash-Lite); $25 (Gemini 2.5 Pro); $5 (Gemini 2.5 Flash); $1.50 (Gemini 2.5 Flash-Lite); open models from $0.28/1M (Llama 3.2 1B)","inferencePricing":"Tuned Gemini 3 and later models serve at 1.5x the base price; older tuned Gemini models at the base price","freeTier":null,"features":["Supervised fine-tuning with adapter sizes from 1 to 16","Preference tuning on Gemini 2.5 Flash and Flash-Lite","Reinforcement learning fine-tuning on Gemini 3.5 Flash","Tuning checkpoints and continuous tuning"],"url":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/tuning","notes":"Now branded Gemini Enterprise Agent Platform. PaLM 2 and Codey tuning went away with those models. The Gemini 2.5 targets retire on Google Cloud on 2026-10-20, so new tunes should start on Gemini 3.5 Flash or 3.1 Flash-Lite. Also a first-party path for tuning open models such as Gemma 3, Llama, and Qwen3."}
+{"id":"mistral-finetune","name":"Mistral Fine-Tuning","vendor":"Mistral","type":"first-party","baseModels":["Mistral text and vision models (legacy fine-tuning API)","Classifier Factory classifiers"],"methods":["lora","full"],"trainingPricing":"$4 minimum per job plus $2/month storage per fine-tuned model; per-token rates on the Mistral pricing page","inferencePricing":"See the Mistral pricing page","freeTier":null,"features":["Supervised fine-tuning for text and vision","Classifier Factory for custom classifiers","European data residency"],"url":"https://docs.mistral.ai/capabilities/finetuning/text-vision-finetuning","notes":"Mistral now files its fine-tuning API under deprecated features and says it is no longer actively supported. Jobs still carry the $4 minimum and $2 monthly storage fee. Fine for keeping an existing tune alive, not a place to start new work."}
+{"id":"together-finetune","name":"Together AI Fine-Tuning","vendor":"Together AI","type":"hosted","baseModels":["Qwen3.5, Qwen3.6, and Qwen3.8 27B","DeepSeek V4 Flash 0731 and DeepSeek-V3.1","Kimi K2.6 and K2.7 Code","Llama 3.x, Llama 4 Scout and Maverick","Gemma 4","gpt-oss-20B and gpt-oss-120B"],"methods":["lora","full","dpo"],"trainingPricing":"LoRA SFT per 1M tokens: $0.34 (models up to 9B), $1.05 (27B to 35B), $2.03 (Llama 3.3 70B), $2.50 (gpt-oss-120B), $6.00 (DeepSeek V4 Flash 0731), $15.00 (Kimi K2.6). DPO about 2.5x SFT. $4 minimum per job, rising to $60 on the largest models.","inferencePricing":"Serve tuned adapters on Together serverless or dedicated endpoints; see Together inference pricing","freeTier":null,"features":["LoRA and full fine-tuning","SFT and DPO","Published per-model prices and minimums","Frontier-scale open models including Kimi K2.6 and DeepSeek V4 Flash"],"url":"https://www.together.ai/pricing","notes":"A broad hosted menu of open models to tune, from Qwen3.5 0.8B up to Kimi K2.6, with a published price and minimum charge for every one. The default for teams that want a custom open model without writing a training loop."}
+{"id":"fireworks-finetune","name":"Fireworks Fine-Tuning","vendor":"Fireworks AI","type":"hosted","baseModels":["Qwen3 family including Qwen3-235B","gpt-oss-120B","DeepSeek V3 family","Kimi K2"],"methods":["lora","full","dpo","rlhf"],"trainingPricing":"Per 1M training tokens, LoRA SFT: $0.50 (up to 16B), $3 (16B to 80B), $6 (80B to 300B), $10 (over 300B). LoRA DPO and full-parameter SFT cost double; full-parameter DPO four times. Reinforcement fine-tuning billed per GPU hour.","inferencePricing":"Fine-tuned models serve at the same price as their base models","freeTier":null,"features":["LoRA and full-parameter SFT and DPO","Reinforcement fine-tuning at on-demand GPU rates","Multi-LoRA serving at base-model prices"],"url":"https://fireworks.ai/pricing","notes":"Clear size-tiered pricing and no serving markup on tuned models, so many variants can share one base cheaply. No longer LoRA-only: full-parameter training and RFT are both on the price list."}
+{"id":"thinking-machines-tinker","name":"Tinker","vendor":"Thinking Machines Lab","type":"hosted","baseModels":["Inkling and Inkling-Small","GLM-5.3","Kimi-K2.6","Nemotron 3 Nano, Super, Ultra and Nemotron 3.5 Lightning","Qwen3.5, Qwen3.6, and Qwen3.8-27B","DeepSeek-V3.1","gpt-oss-20B and gpt-oss-120B"],"methods":["lora","rlhf"],"trainingPricing":"Per 1M training tokens: $0.44 (Qwen3-8B), $0.737 (gpt-oss-120B), $3.718 (DeepSeek-V3.1), $4.84 (Kimi-K2.6), $5.61 (Inkling at a limited-time 50% discount). Prefill and sampling billed separately.","inferencePricing":"Sampling from checkpoints billed per token; serverless inference in beta for Inkling models only","freeTier":null,"features":["Low-level API: forward_backward, optim_step, sample, save_state","Built-in RL losses (PPO, CISPO, DRO) and custom losses","Downloadable LoRA checkpoints, stored at $0.10/GB-month","Priced on active parameters, which favors MoE models"],"url":"https://tinker-docs.thinkingmachines.ai/tinker/models/","notes":"A research-grade training API rather than a job runner: you write the training loop and Thinking Machines runs the GPUs. Checkpoints can be downloaded, and pricing follows active parameters, which makes trillion-parameter MoE models such as Kimi-K2.6 practical to tune."}
+{"id":"openpipe","name":"OpenPipe (CoreWeave)","vendor":"CoreWeave","type":"hosted","baseModels":["Open-weight models through managed SFT","Bring-your-own models through the ART reinforcement learning library"],"methods":["lora","rlhf"],"trainingPricing":"Not published; confirm with CoreWeave","inferencePricing":"Not published; confirm with CoreWeave","freeTier":null,"features":["Agent Reinforcement Trainer (ART) open-source RL library","Managed SFT workflows","Sits alongside Weights & Biases inside CoreWeave"],"url":"https://openpipe.ai","notes":"CoreWeave agreed to acquire OpenPipe in September 2025. OpenPipe said it would keep supporting existing customers, the ART library, and managed SFT, while warning the product may evolve. The old distill-from-GPT-4-traces pitch is now part of a larger CoreWeave training stack, so older published pricing should not be relied on."}
+{"id":"predibase","name":"Predibase (Rubrik)","vendor":"Rubrik","type":"hosted","baseModels":["Open-weight models (legacy Predibase platform)"],"methods":["lora","qlora","full"],"trainingPricing":"Not published","inferencePricing":"Not published","freeTier":null,"features":["Multi-adapter LoRA serving","LoRAX open-source serving project still maintained on GitHub"],"url":"https://github.com/predibase/lorax","notes":"Rubrik agreed to acquire Predibase in June 2025, and predibase.com now redirects to Rubrik Agent Cloud. The standalone self-serve fine-tuning product is no longer marketed on its own; LoRAX, the open-source multi-LoRA server, is the piece most teams can still use directly."}
+{"id":"huggingface-autotrain","name":"Hugging Face AutoTrain","vendor":"Hugging Face","type":"hosted","baseModels":["Most HuggingFace open-weights models"],"methods":["lora","full","dpo"],"trainingPricing":"GPU time on the Hugging Face Space hardware it runs on","inferencePricing":"Inference Endpoints billed separately by instance hour","freeTier":"Free to run locally or on your own hardware","features":["No-code UI in a Docker Space","Local UI and Python API","Push to Hub"],"url":"https://huggingface.co/docs/autotrain/index","notes":"Hugging Face now marks AutoTrain as no longer maintained, with no new features or bug fixes, and recommends Axolotl, TRL, or transformers Trainer instead. NVIDIA agreed on 2026-09-03 to acquire Hugging Face for $12.93B and says the Hub will stay open and accelerator-neutral."}
+{"id":"aws-bedrock-finetune","name":"AWS Bedrock Fine-Tuning","vendor":"Amazon Web Services","type":"hosted","baseModels":["Amazon Nova 2 Lite, Nova Pro, Nova Lite, Nova Micro","Llama 3.1, 3.2, and 3.3","gpt-oss-20B and Qwen3 32B (reinforcement fine-tuning)","Nova Canvas and Titan image and embedding models"],"methods":["lora","rlhf","continued-pretraining"],"trainingPricing":"Per-model rates on the Bedrock pricing page","inferencePricing":"Customized Nova models run on demand at the base model price","freeTier":null,"features":["Supervised fine-tuning and distillation","Reinforcement fine-tuning with GRPO","Continued pre-training","Nova Forge subscription for deeper customization"],"url":"https://docs.aws.amazon.com/bedrock/latest/userguide/custom-model-fine-tuning.html","notes":"Nova is now the center of Bedrock customization, and on-demand serving for tuned Nova models removes the old Provisioned Throughput cost at low volume. The Claude path is gone: Claude 3 Haiku was the last customizable Claude model and reached Bedrock end of life on 2026-09-10."}
+{"id":"replicate-finetune","name":"Replicate Fine-Tuning","vendor":"Replicate (Cloudflare)","type":"hosted","baseModels":["FLUX image models","SDXL","Community trainers for open language models"],"methods":["lora"],"trainingPricing":"GPU time; for example H100 at $0.001525/sec","inferencePricing":"Per-second GPU billing; fast-booting fine-tunes bill only active time","freeTier":null,"features":["Image LoRA fine-tuning standout","API-driven","Fast-booting fine-tunes without idle charges"],"url":"https://replicate.com/pricing","notes":"Strongest image LoRA experience (FLUX, SDXL). Cloudflare agreed to acquire Replicate in November 2025 to fold its model catalog into Workers. Less competitive for text-LLM fine-tuning than Together or Fireworks."}
+{"id":"modal-finetune","name":"Modal Labs (DIY)","vendor":"Modal","type":"hosted","baseModels":["Any (BYO code)","Open-model recipes published"],"methods":["lora","qlora","full","dpo","rlhf","continued-pretraining"],"trainingPricing":"GPU-time billing per second; H100 about $3.95/hr, B200 about $6.25/hr","inferencePricing":"GPU-time billing on serving function","freeTier":"$30/mo compute credit on Starter","features":["Full BYO control","Serverless GPU","Distributed training","Open-source recipes"],"url":"https://modal.com/docs/examples/llm-finetuning","notes":"BYO fine-tuning infra. You own the code; Modal owns the GPU scheduling. The right answer if your tuning recipe does not fit a turn-key provider."}

2026-10-12/frameworks.jsonl ADDED Viewed

	@@ -0,0 +1,17 @@

+{"id":"langchain","name":"LangChain","vendor":"LangChain Inc.","languages":["python","typescript"],"version":"1.4.x","released":"2022-10-22","license":"MIT","githubStarsK":146.3,"weeklyInstallsK":39683,"category":"agent-orchestration","features":["chains","tools","memory","tracing (LangSmith)","document loaders","500+ integrations"],"url":"https://www.langchain.com","github":"https://github.com/langchain-ai/langchain","notes":"The most-installed Python agent framework, now on the 1.x line and billed by its maintainers as \"the agent engineering platform.\" langchain-core and per-provider packages ship separately to reduce churn. Pairs tightly with LangGraph for stateful multi-step workflows and LangSmith for tracing."}
+{"id":"langgraph","name":"LangGraph","vendor":"LangChain Inc.","languages":["python","typescript"],"version":"1.2.x","released":"2024-01-17","license":"MIT","githubStarsK":41.6,"weeklyInstallsK":10044,"category":"agent-orchestration","features":["stateful graphs","checkpointing","human-in-the-loop","streaming","time-travel"],"url":"https://docs.langchain.com/oss/python/langgraph/overview","github":"https://github.com/langchain-ai/langgraph","notes":"Stateful agent graphs from the LangChain team, now past 1.0. The framework many production agents moved to because of first-class checkpointing, human-in-the-loop interrupts, and time-travel debugging. Draws about 10 million weekly PyPI installs as of September 2026."}
+{"id":"llamaindex","name":"LlamaIndex","vendor":"LlamaIndex Inc.","languages":["python","typescript"],"version":"0.14.x","released":"2022-11-22","license":"MIT","githubStarsK":52.2,"weeklyInstallsK":708,"category":"rag","features":["document loaders","indexes","query engines","agents","workflows","observability"],"url":"https://www.llamaindex.ai","github":"https://github.com/run-llama/llama_index","notes":"RAG-first framework, strong on document ingestion, indexing strategies, and query rewriting. The company now describes LlamaIndex as a document processing platform for AI, with LlamaParse and hosted services alongside the open-source library. Workflows is the agent-orchestration layer."}
+{"id":"microsoft-agent-framework","name":"Microsoft Agent Framework","vendor":"Microsoft","languages":["python","multi"],"version":"1.18.x","released":"2025-10-01","license":"MIT","githubStarsK":13.5,"weeklyInstallsK":37,"category":"multi-agent","features":["graph-based workflows","sequential, concurrent, handoff, and group patterns","A2A + MCP","durable, restartable runs","human-in-the-loop","Python + .NET + Go"],"url":"https://learn.microsoft.com/en-us/agent-framework/","github":"https://github.com/microsoft/agent-framework","notes":"Microsoft's successor to both AutoGen and Semantic Kernel, each of which ships a migration guide to it. Public preview landed October 1, 2025 and 1.0 shipped for Python and .NET on April 2, 2026, with a Go SDK in a separate repo. The default Microsoft pick for new multi-agent projects; weekly figure is PyPI agent-framework only and excludes NuGet."}
+{"id":"autogen","name":"AutoGen","vendor":"Microsoft","languages":["python"],"version":"0.7.x (final)","released":"2023-09-25","license":"CC-BY-4.0","githubStarsK":61,"weeklyInstallsK":87,"category":"multi-agent","features":["conversable agents","group chat","code execution","human approval","event-driven"],"url":"https://microsoft.github.io/autogen/","github":"https://github.com/microsoft/autogen","notes":"Microsoft Research multi-agent framework, now in maintenance mode: no new features, community managed, and limited to bug fixes and security patches. The last release was python-v0.7.5 on September 30, 2025. Microsoft points new users to Microsoft Agent Framework; listed for its large existing user base."}
+{"id":"crewai","name":"CrewAI","vendor":"crewAI Inc.","languages":["python"],"version":"1.15.x","released":"2023-12-12","license":"MIT","githubStarsK":58.5,"weeklyInstallsK":585,"category":"multi-agent","features":["roles","tasks","crews","flows","tool registry"],"url":"https://crewai.com","github":"https://github.com/crewAIInc/crewAI","notes":"Role-based multi-agent framework. The \"Agent + Task + Crew\" abstraction maps cleanly onto how teams actually think about delegation. Hugely popular for marketing, research, and sales agent setups."}
+{"id":"pydantic-ai","name":"Pydantic AI","vendor":"Pydantic","languages":["python"],"version":"2.43.x","released":"2024-12-02","license":"MIT","githubStarsK":19.9,"weeklyInstallsK":1222,"category":"agent-orchestration","features":["type-safe agents","pydantic models","streaming","logfire integration","function calling","realtime voice"],"url":"https://pydantic.dev/pydantic-ai","github":"https://github.com/pydantic/pydantic-ai","notes":"Type-safe agent framework from the Pydantic team, now on its 2.x line and pitched as typed end to end across agents, realtime voice, image generation, and embeddings. Strong fit for Python codebases that already use Pydantic for data validation. Draws about 1.2 million weekly PyPI installs as of September 2026."}
+{"id":"mastra","name":"Mastra","vendor":"Mastra","languages":["typescript"],"version":"1.66.x","released":"2024-09-12","license":"Apache-2.0 (ee/ directories under a separate enterprise license)","githubStarsK":28,"weeklyInstallsK":1240,"category":"agent-orchestration","features":["workflows","agents","rag","evals","memory","voice"],"url":"https://mastra.ai","github":"https://github.com/mastra-ai/mastra","notes":"TypeScript-first agent framework, built by the team behind Gatsby. Targets Node.js and Next.js stacks where most agent libraries are Python-only. Strong DX with first-class evals. Licensed Apache-2.0, with enterprise features kept under ee/ directories. Weekly figure is npm @mastra/core."}
+{"id":"openai-agents-sdk","name":"OpenAI Agents SDK","vendor":"OpenAI","languages":["python","typescript"],"version":"0.22.x","released":"2025-03-12","license":"MIT","githubStarsK":29.4,"weeklyInstallsK":2784,"category":"sdk","features":["handoffs","guardrails","tracing","sessions","voice agents","sandbox agents"],"url":"https://openai.github.io/openai-agents-python/","github":"https://github.com/openai/openai-agents-python","notes":"OpenAI's official client-side agent SDK. Handoffs are first-class for multi-agent, and it integrates tightly with OpenAI tracing. For a fully hosted alternative, OpenAI's Agents API (public beta since September 10, 2026) runs the Codex harness server-side. Use the SDK if you want the loop in your own process; otherwise LangGraph or Pydantic AI are usually a better multi-provider fit."}
+{"id":"claude-agent-sdk","name":"Claude Agent SDK","vendor":"Anthropic","languages":["python","typescript"],"version":"0.2.x (Python), 0.3.x (TypeScript)","released":"2025-09-29","license":"MIT","githubStarsK":8.1,"weeklyInstallsK":6977,"category":"sdk","features":["claude code core","mcp servers","hooks","subagents","slash commands"],"url":"https://code.claude.com/docs/en/agent-sdk/overview","github":"https://github.com/anthropics/claude-agent-sdk-python","notes":"The Claude Code core, repackaged as a programmable SDK. Native MCP, hooks, and subagent orchestration. Best fit if your agent needs the same surface area as Claude Code. Weekly figure is PyPI only; the TypeScript package drew about 9 million npm downloads the same week."}
+{"id":"vercel-ai-sdk","name":"Vercel AI SDK","vendor":"Vercel","languages":["typescript","javascript"],"version":"7.x","released":"2023-04-21","license":"Apache-2.0","githubStarsK":26.7,"weeklyInstallsK":18340,"category":"sdk","features":["streaming UI","tool calling","structured output","multi-provider","react/svelte/vue hooks","agents"],"url":"https://ai-sdk.dev","github":"https://github.com/vercel/ai","notes":"TypeScript-first AI SDK, now on v7 with its docs at ai-sdk.dev. Best in class for streaming UI in Next.js and React apps, with agent loops and multi-step tool calling built in. The most-used AI library in JS-land by npm downloads."}
+{"id":"agno","name":"Agno","vendor":"Agno","languages":["python"],"version":"3.x","released":"2024-08-15","license":"Apache-2.0","githubStarsK":42.2,"weeklyInstallsK":393,"category":"agent-orchestration","features":["multi-modal agents","memory","knowledge bases","workflows","monitoring"],"url":"https://www.agno.com","github":"https://github.com/agno-agi/agno","notes":"Phidata rebranded, now on its 3.x line under Apache-2.0 and positioned as a way to build, run, and manage agent platforms. Lightweight Python framework with first-class multimodal (image, audio, video) agents and strong observability through Agno Platform."}
+{"id":"smolagents","name":"smolagents","vendor":"Hugging Face","languages":["python"],"version":"1.26.x","released":"2024-12-31","license":"Apache-2.0","githubStarsK":29.3,"weeklyInstallsK":93,"category":"agent-orchestration","features":["code agents","hub integration","multi-step","minimal abstractions"],"url":"https://huggingface.co/docs/smolagents","github":"https://github.com/huggingface/smolagents","notes":"Minimal agent framework from Hugging Face. CodeAgent expresses tool use as Python code instead of JSON, which often produces better trajectories on agentic benchmarks. Releases have slowed (1.26.0 shipped May 29, 2026), though the repo is still maintained."}
+{"id":"haystack","name":"Haystack","vendor":"deepset","languages":["python"],"version":"3.1.x","released":"2020-04-13","license":"Apache-2.0","githubStarsK":26.5,"weeklyInstallsK":128,"category":"rag","features":["pipelines","document stores","agents","evaluation","hybrid retrieval"],"url":"https://haystack.deepset.ai","github":"https://github.com/deepset-ai/haystack","notes":"One of the original RAG frameworks (predates the LLM era), now on the 3.x line. Modular pipelines and agent workflows with explicit control over retrieval, routing, and memory. Strong for enterprise RAG and competitive with LangGraph for retrieval-heavy workflows."}
+{"id":"browser-use","name":"browser-use","vendor":"Browser Use","languages":["python"],"version":"0.13.x","released":"2024-11-15","license":"MIT","githubStarsK":114.6,"weeklyInstallsK":1651,"category":"browser-agent","features":["playwright","vision","click/type/scroll","multi-tab","file download"],"url":"https://browser-use.com","github":"https://github.com/browser-use/browser-use","notes":"Open-source browser automation framework for AI agents. Passes pages to LLMs as DOM plus screenshots so the model can decide each action. Past 110,000 GitHub stars as of September 2026."}
+{"id":"google-adk","name":"Agent Development Kit (ADK)","vendor":"Google","languages":["python","typescript","multi"],"version":"2.9.x","released":"2025-04-09","license":"Apache-2.0","githubStarsK":21.5,"weeklyInstallsK":2262,"category":"agent-orchestration","features":["code-first agents","graph workflows","multi-agent","evaluation","deployment tooling","Python + TypeScript + Go + Java + Kotlin"],"url":"https://adk.dev","github":"https://github.com/google/adk-python","notes":"Google's open-source, code-first framework for building, evaluating, and deploying agents. Optimized for Gemini but model-agnostic and deployment-agnostic. ADK 2.0 brought graph workflows, and SDKs now span Python, TypeScript, Go, Java, and Kotlin. Weekly figure is PyPI google-adk only."}
+{"id":"pipecat","name":"Pipecat","vendor":"Daily","languages":["python"],"version":"1.10.x","released":"2024-04-01","license":"BSD-2-Clause","githubStarsK":15.5,"weeklyInstallsK":247,"category":"voice-agent","features":["STT","TTS","WebRTC","turn detection","multi-modal"],"url":"https://pipecat.ai","github":"https://github.com/pipecat-ai/pipecat","notes":"Real-time voice agent framework maintained by Daily and the community, now past 1.0. Pipelines connect STT (Deepgram, Whisper), LLMs, and TTS (Cartesia, ElevenLabs) over WebRTC with sub-second turn-taking. The default OSS stack for production voice agents."}

2026-10-12/funding.jsonl ADDED Viewed

	@@ -0,0 +1,43 @@

+{"id":"instinct-2026-09","company":"Instinct","category":"agent","stage":"series-c","amountM":1000,"valuationB":10,"announcedDate":"2026-09-28","leadInvestors":[],"notableInvestors":["Sequoia Capital","Benchmark","Coatue"],"description":"Personal AI agent for everyday tasks that users text or call; it works through its own phone and computer and is still in early access. The Series C came about a month after a $250M Series B at $2.5B. The release names the three investors without designating a lead and does not say whether the $10B valuation is pre-money or post-money.","url":"https://instinct.com","sourceUrl":"https://www.businesswire.com/news/home/20260928153437/en/Instinct-Raises-%241-Billion-in-Series-C-Funding-from-Sequoia-Benchmark-and-Coatue-at-%2410-Billion-Valuation"}
+{"id":"snorkel-2026-09","company":"Snorkel AI","category":"data","stage":"series-e","amountM":350,"valuationB":3.5,"announcedDate":"2026-09-22","leadInvestors":["Insight Partners","S32"],"notableInvestors":["March Capital","Blumberg Capital","Allegis Capital","Third Point Ventures","Greylock","Lightspeed","GV","Prosperity7 Ventures","Wells Fargo"],"description":"Training data for frontier labs. Snorkel shifted from data-labeling software to delivering finished expert datasets as a service; the round values it at $3.5B, close to triple the $1.3B mark set by its $100M Series D in 2025.","url":"https://snorkel.ai","sourceUrl":"https://www.prnewswire.com/news-releases/snorkel-ai-raises-350m-to-scale-the-data-factory-for-frontier-ai-302886796.html"}
+{"id":"crusoe-2026-09","company":"Crusoe","category":"infra","stage":"series-f","amountM":3900,"valuationB":30.9,"announcedDate":"2026-09-17","leadInvestors":["Atreides Management","Mubadala Capital","Valor Equity Partners"],"notableInvestors":["Founders Fund","GIC","NVIDIA","Qatar Investment Authority","Radical Ventures","TPG","Altimeter","Fidelity Management & Research","Salesforce Ventures","T. Rowe Price","Tiger Global"],"description":"Vertically integrated AI infrastructure provider spanning power, data centers and Crusoe Cloud. Initial closing of a $3.9B Series F at a $30.9B post-money valuation; the company reports more than $140B in total contracted value and over 6 GW of gross contracted capacity, 1 GW of it delivered and operational.","url":"https://crusoe.ai","sourceUrl":"https://www.crusoe.ai/resources/newsroom/crusoe-announces-series-f-funding"}
+{"id":"factory-2026-09","company":"Factory","category":"coding","stage":"growth","amountM":200,"valuationB":5,"announcedDate":"2026-09-15","leadInvestors":["Blackstone","Khosla Ventures"],"notableInvestors":["Sequoia Capital","Insight Partners","Evantic Capital","Sound Ventures","New Enterprise Associates","Clearlake"],"description":"Enterprise platform for autonomous software development: model selection, deployment and measurement across the whole development lifecycle rather than an editor. Third large raise in under a year, taking total funding past $400M.","url":"https://factory.com","sourceUrl":"https://factory.com/news/5-billion-valuation"}
+{"id":"temporal-2026-09","company":"Temporal Technologies","category":"infra","stage":"series-e","amountM":550,"valuationB":12.55,"announcedDate":"2026-09-14","leadInvestors":["Lightspeed Venture Partners","Wellington Management"],"notableInvestors":["Goldman Sachs Alternatives","Tiger Global","T. Rowe Price","SV Angel","Andreessen Horowitz","Sequoia Capital","Index Ventures","GIC","Sapphire Ventures"],"description":"Durable execution platform that keeps long-running agent and application workflows from losing state across failures. Temporal says run-rate revenue grew over 200% year over year, Temporal Cloud processed 1.9 trillion billable actions in August 2026, and paying customers passed 4,300 including OpenAI, Snap and NVIDIA.","url":"https://temporal.io","sourceUrl":"https://temporal.io/blog/temporal-raises-usd550m-series-e-at-usd12-55b-valuation-ai"}
+{"id":"positron-2026-09","company":"Positron AI","category":"inference","stage":"series-c","amountM":875,"valuationB":5,"announcedDate":"2026-09-10","leadInvestors":["New Enterprise Associates","Andra Capital","Atreides Management","Valor Equity Partners","SemiAnalysis Capital"],"notableInvestors":["Jim Clark","DFJ Growth","Qatar Investment Authority","Hudson River Trading","Cisco Investments","Naver Ventures"],"description":"Inference silicon designed around memory capacity rather than peak FLOPS. Two tranches: a $375M Series C at $3.5B pre-money plus a Series C-1 of up to $500M led by NEA and Jim Clark, for $875M at $5B post-money. Funds the tapeout of its next chip, Asimov, and the production ramp of the Titan system.","url":"https://www.positron.ai","sourceUrl":"https://www.prnewswire.com/news-releases/positron-ai-raises-875-million-at-a-5-billion-valuation-to-bring-its-next-generation-inference-silicon-to-market-302874601.html"}
+{"id":"harvey-2026-09","company":"Harvey","category":"enterprise","stage":"growth","amountM":550,"valuationB":15.5,"announcedDate":"2026-09-09","leadInvestors":["Diffusion","Lightspeed Venture Partners"],"notableInvestors":["Sequoia Capital","Kleiner Perkins","Andreessen Horowitz","Coatue","GIC","Goldman Sachs Alternatives","Sapphire Ventures","Whale Rock"],"description":"Legal AI. Raised at $15.5B less than six months after a $200M round at $11B in March 2026; Harvey says 80% of Am Law 100 firms and five of the Fortune 10 in-house legal teams use it. Sapphire Ventures and Whale Rock came in as new investors.","url":"https://www.harvey.ai","sourceUrl":"https://www.harvey.ai/blog/harvey-raises-dollar550m-at-a-dollar155b-valuation-to-help-legal-teams-own-their-intelligence"}
+{"id":"cognition-2026-09","company":"Cognition","category":"coding","stage":"series-e","amountM":2000,"valuationB":48,"announcedDate":"2026-09-08","leadInvestors":["Andreessen Horowitz","Accel"],"notableInvestors":["Founders Fund","General Catalyst","Avenir","NVIDIA","Benchmark","Bessemer Venture Partners","Kleiner Perkins","Greylock","Lightspeed","Altimeter Capital","Bond Capital"],"description":"Maker of the Devin coding agent. Over $2B at $48B post-money, nearly double the $26B Series D price set three months earlier; Cognition says run-rate revenue went from $492M to almost $900M over that stretch.","url":"https://cognition.ai","sourceUrl":"https://cognition.com/blog/series-e"}
+{"id":"mistral-2026-09","company":"Mistral AI","category":"frontier-lab","stage":"series-d","amountM":3500,"valuationB":24,"announcedDate":"2026-09-08","leadInvestors":["Samsung Electronics"],"notableInvestors":["Scaleup Europe Fund (EQT)","PSG Equity","ASML","NVIDIA","BlackRock","Andreessen Horowitz"],"description":"EUR 3B Series D at a post-money valuation above EUR 21B (about $24B), led by Samsung Electronics with the EQT-managed Scaleup Europe Fund and PSG Equity co-leading. Mistral calls it the largest equity round ever raised by a European tech company; the money goes to owned data centers and compute.","url":"https://mistral.ai","sourceUrl":"https://mistral.ai/news/mistral-makes-sovereign-open-weight-ai-to-frontier/"}
+{"id":"databricks-2026-08","company":"Databricks","category":"data","stage":"growth","amountM":5000,"valuationB":190,"announcedDate":"2026-08-13","leadInvestors":["Coatue"],"notableInvestors":["Blackstone","MGX","T. Rowe Price","Sixth Street Growth","TPG","Point72"],"description":"Second $5B raise of 2026, at $190B (up from $134B in February). Databricks crossed a $7B revenue run-rate growing over 80% year over year; proceeds fund Lakebase, Genie and Unity AI Gateway for enterprise agents.","url":"https://www.databricks.com","sourceUrl":"https://www.databricks.com/company/newsroom/press-releases/databricks-grows-80-yoy-surpasses-7b-revenue-run-rate-scales"}
+{"id":"fireworks-2026-07","company":"Fireworks AI","category":"inference","stage":"series-d","amountM":1505,"valuationB":17.5,"announcedDate":"2026-07-16","leadInvestors":["Atreides Management","Index Ventures","TCV"],"notableInvestors":["Lightspeed Venture Partners","NVIDIA","Bessemer Venture Partners","Menlo Ventures"],"description":"Inference and fine-tuning platform for open models. Raised at $17.5B, up from a $4B valuation in October 2025, after passing $1B in annualized revenue and more than 40 trillion tokens served per day.","url":"https://fireworks.ai","sourceUrl":"https://fireworks.ai/blog/series-d-announcement"}
+{"id":"sambanova-2026-07","company":"SambaNova","category":"inference","stage":"series-f","amountM":1000,"valuationB":11,"announcedDate":"2026-07-08","leadInvestors":["General Atlantic"],"notableInvestors":["Seligman Ventures","T. Rowe Price","Capital Group","Intel Capital","BlackRock","Qatar Investment Authority"],"description":"First close of a Series F for the RDU inference-chip and systems maker, five months after a $350M Series E. JPMorganChase was announced the same day as an on-prem inference customer.","url":"https://sambanova.ai","sourceUrl":"https://www.generalatlantic.com/media-article/sambanova-completes-first-close-of-1-billion-financing-at-11-billion-valuation/"}
+{"id":"kling-2026-07","company":"Kling AI","category":"video","stage":"growth","amountM":2800,"valuationB":17.8,"announcedDate":"2026-07-02","leadInvestors":["Alibaba","Tencent"],"notableInvestors":["Baidu"],"description":"Kuaishou spun out its AI video generator with over CNY 19B (about $2.8B) of committed capital at a $15B pre-money valuation, per Kuaishou's Hong Kong exchange filing. The round is capped at CNY 20.45B and dilutes Kuaishou to about 68%.","url":"https://klingai.com","sourceUrl":"https://www.reuters.com/world/china/alibaba-tencent-back-kuaishous-kling-ai-28-billion-fundraise-2026-07-03/"}
+{"id":"together-2026-07","company":"Together AI","category":"inference","stage":"series-c","amountM":800,"valuationB":8.3,"announcedDate":"2026-07-01","leadInvestors":["Aramco Ventures"],"notableInvestors":["Vista Equity Partners","General Catalyst","NVIDIA","Emergence Capital"],"description":"Open-model inference cloud. Series C more than doubles the $3.3B Series B mark; Together says new investors separately lined up commitments for more than 500 MW of compute capacity.","url":"https://www.together.ai","sourceUrl":"https://www.together.ai/blog/announcing-our-series-c"}
+{"id":"baseten-2026-06","company":"Baseten","category":"inference","stage":"series-f","amountM":1500,"valuationB":13,"announcedDate":"2026-06-22","leadInvestors":["Altimeter Capital","Conviction","Spark Capital"],"notableInvestors":["Sands Capital","Wellington Management","IVP","Greylock","Blackbird"],"description":"Inference platform, fourth raise in 18 months. Split-priced round: the Wall Street Journal reported some investors came in at $13B and others at $11B. Five months earlier it raised a $300M Series E at $5B.","url":"https://www.baseten.co","sourceUrl":"https://www.baseten.co/blog/announcing-our-series-f/"}
+{"id":"deepseek-2026-06","company":"DeepSeek","category":"frontier-lab","stage":"growth","amountM":7400,"valuationB":51.8,"announcedDate":"2026-06-16","leadInvestors":["Tencent","CATL"],"notableInvestors":["China National AI Industry Investment Fund","NetEase","JD.com"],"description":"First external round, about CNY 50B, never announced by the company. Founder Liang Wenfeng committed CNY 20B; outside capital went through a founder-managed partnership with a five-year lock-up. A July 16 Shenzhen filing by Anhui Korrun implied a CNY 350.88B (about $51.8B) post-money value.","url":"https://www.deepseek.com","sourceUrl":"https://www.reuters.com/world/asia-pacific/chinas-deepseek-closes-over-7-billion-funding-with-unusual-deal-structure-2026-06-16/"}
+{"id":"anthropic-2026-05","company":"Anthropic","category":"frontier-lab","stage":"growth","amountM":65000,"valuationB":965,"announcedDate":"2026-05-28","leadInvestors":["Altimeter Capital","Dragoneer","Greenoaks","Sequoia Capital"],"notableInvestors":["Capital Group","Coatue","D1 Capital Partners","GIC","ICONIQ","XN","Jane Street","Amazon"],"description":"Series H at $965B post-money, above OpenAI's $852B March mark. Co-led by Capital Group, Coatue, D1, GIC, ICONIQ and XN; the total includes $15B of previously committed hyperscaler investment, including $5B from Amazon. Run-rate revenue crossed $47B in May.","url":"https://www.anthropic.com","sourceUrl":"https://www.anthropic.com/news/series-h"}
+{"id":"cognition-2026-05","company":"Cognition","category":"coding","stage":"series-d","amountM":1000,"valuationB":26,"announcedDate":"2026-05-27","leadInvestors":["Lux Capital","General Catalyst","8VC"],"notableInvestors":["Founders Fund","Ribbit Capital","Atreides Management","Elad Gil"],"description":"Maker of the Devin coding agent (and Windsurf, acquired in 2025). Over $1B at $25B pre-money, up from $10.2B in September 2025, on $492M of annualized revenue. Superseded three months later by the Series E at $48B.","url":"https://cognition.ai","sourceUrl":"https://cognition.com/blog/series-d"}
+{"id":"moonshot-2026-05","company":"Moonshot AI","category":"frontier-lab","stage":"growth","amountM":2000,"valuationB":20,"announcedDate":"2026-05-07","leadInvestors":["Meituan Long-Z Investments"],"notableInvestors":["Tsinghua Capital","China Mobile","CPE Yuanfeng"],"description":"Beijing lab behind the open-weight Kimi models. About $2B at a valuation above $20B, per adviser Huafeng Capital; ARR topped $200M in April and Moonshot raised about $3.9B in six months.","url":"https://www.moonshot.ai","sourceUrl":"https://www.bloomberg.com/news/articles/2026-05-07/kimi-chatbot-maker-moonshot-ai-valued-at-20-billion-in-meituan-led-round"}
+{"id":"sierra-2026-05","company":"Sierra","category":"agent","stage":"growth","amountM":950,"valuationB":15.8,"announcedDate":"2026-05-04","leadInvestors":["Tiger Global","GV"],"notableInvestors":["Benchmark","Sequoia Capital","Greenoaks"],"description":"Customer-experience agent platform co-founded by Bret Taylor and Clay Bavor. Sierra put the valuation at over $15B (CNBC reported $15.8B post-money) and says it serves more than 40% of the Fortune 50.","url":"https://sierra.ai","sourceUrl":"https://sierra.ai/blog/better-customer-experiences-built-on-sierra"}
+{"id":"anthropic-google-2026-04","company":"Anthropic","category":"frontier-lab","stage":"strategic","amountM":10000,"valuationB":380,"announcedDate":"2026-04-24","leadInvestors":["Google"],"notableInvestors":[],"description":"Google put in $10B of cash at the Series G price ($350B pre-money, $380B post) and committed up to $30B more contingent on performance milestones. Follows the April deal with Google and Broadcom for 5 GW of TPU capacity starting in 2027.","url":"https://www.anthropic.com","sourceUrl":"https://www.cnbc.com/2026/04/24/google-to-invest-up-to-40-billion-in-anthropic-as-search-giant-spreads-its-ai-bets.html"}
+{"id":"anthropic-amazon-2026-04","company":"Anthropic","category":"frontier-lab","stage":"strategic","amountM":5000,"valuationB":null,"announcedDate":"2026-04-20","leadInvestors":["Amazon"],"notableInvestors":[],"description":"Amazon invested $5B and committed up to $20B more tied to commercial milestones, on top of $8B invested since 2023. Anthropic committed more than $100B of AWS spend over ten years for up to 5 GW, including Trainium2 through Trainium4 capacity.","url":"https://www.anthropic.com","sourceUrl":"https://www.aboutamazon.com/news/company-news/amazon-invests-additional-5-billion-anthropic-ai"}
+{"id":"openai-2026-03","company":"OpenAI","category":"frontier-lab","stage":"growth","amountM":122000,"valuationB":852,"announcedDate":"2026-03-31","leadInvestors":["SoftBank","Andreessen Horowitz","D. E. Shaw Ventures","MGX","TPG","T. Rowe Price"],"notableInvestors":["Amazon","NVIDIA","Microsoft","Thrive Capital","Sequoia Capital"],"description":"Closed with $122B of committed capital at $852B post-money, up from the $110B announced February 27. Anchored by Amazon ($50B), NVIDIA ($30B) and SoftBank ($30B); over $3B came from individuals through bank channels. OpenAI said revenue was running at $2B a month.","url":"https://openai.com","sourceUrl":"https://openai.com/index/accelerating-the-next-phase-ai/"}
+{"id":"anthropic-2026-02","company":"Anthropic","category":"frontier-lab","stage":"growth","amountM":30000,"valuationB":380,"announcedDate":"2026-02-12","leadInvestors":["GIC","Coatue"],"notableInvestors":["D. E. Shaw Ventures","Dragoneer","Founders Fund","ICONIQ","MGX"],"description":"Series G at $380B post-money, co-led by D. E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ and MGX. Includes part of the investments Microsoft and NVIDIA announced in November 2025. Run-rate revenue was $14B at announcement.","url":"https://www.anthropic.com","sourceUrl":"https://www.anthropic.com/news/anthropic-raises-30-billion-series-g-funding-380-billion-post-money-valuation"}
+{"id":"decagon-2026-01","company":"Decagon","category":"agent","stage":"series-d","amountM":250,"valuationB":4.5,"announcedDate":"2026-01-28","leadInvestors":["Coatue","Index Ventures"],"notableInvestors":["Andreessen Horowitz","Accel","Bain Capital Ventures","Ribbit Capital"],"description":"AI customer-service agents for enterprise. Series D tripled the valuation in about six months; Decagon added more than 100 enterprise customers in 2025, including Avis Budget Group and Deutsche Telekom.","url":"https://decagon.ai","sourceUrl":"https://decagon.ai/blog/series-d-announcement"}
+{"id":"openai-2025-10","company":"OpenAI","category":"frontier-lab","stage":"tender","amountM":6600,"valuationB":500,"announcedDate":"2025-10-02","leadInvestors":[],"notableInvestors":["Thrive Capital","SoftBank","Dragoneer","MGX","T. Rowe Price"],"description":"Secondary sale in which current and former employees sold about $6.6B of stock at a $500B valuation, short of the $10.3B authorized. Buyers included Thrive, SoftBank, Dragoneer, MGX and T. Rowe Price.","url":"https://openai.com","sourceUrl":"https://www.cnbc.com/2025/10/02/openai-share-sale-500-billion-valuation.html"}
+{"id":"mistral-2025-09","company":"Mistral AI","category":"frontier-lab","stage":"series-c","amountM":2000,"valuationB":13.8,"announcedDate":"2025-09-09","leadInvestors":["ASML"],"notableInvestors":[],"description":"EUR 1.7B Series C at a EUR 11.7B post-money valuation. ASML put in EUR 1.3B for about 11% and a seat on Mistral's strategic committee. Precedes the Samsung-led Series D of September 2026.","url":"https://mistral.ai","sourceUrl":"https://mistral.ai/news/mistral-ai-raises-1-7-b-to-accelerate-technological-progress-with-ai/"}
+{"id":"sierra-2025-09","company":"Sierra","category":"agent","stage":"growth","amountM":350,"valuationB":10,"announcedDate":"2025-09-04","leadInvestors":["Greenoaks"],"notableInvestors":[],"description":"Conversational agents for customer service, founded by Bret Taylor and Clay Bavor. More than doubled the $4.5B valuation Greenoaks led in October 2024.","url":"https://sierra.ai","sourceUrl":"https://sierra.ai/blog/theres-an-agent-for-that-and-it-runs-on-sierra"}
+{"id":"anthropic-2025-09","company":"Anthropic","category":"frontier-lab","stage":"series-f","amountM":13000,"valuationB":183,"announcedDate":"2025-09-02","leadInvestors":["ICONIQ"],"notableInvestors":["Fidelity Management & Research","Lightspeed Venture Partners","Qatar Investment Authority","Blackstone","Coatue"],"description":"Series F at $183B post-money, co-led by Fidelity and Lightspeed. Up from $61.5B in March 2025; predates the 2026 Series G and Series H.","url":"https://www.anthropic.com","sourceUrl":"https://www.anthropic.com/news/anthropic-raises-series-f-at-usd183b-post-money-valuation"}
+{"id":"cohere-2025-08","company":"Cohere","category":"frontier-lab","stage":"growth","amountM":500,"valuationB":6.8,"announcedDate":"2025-08-14","leadInvestors":["Radical Ventures","Inovia Capital"],"notableInvestors":["AMD Ventures","NVIDIA","PSP Investments","Salesforce Ventures"],"description":"Enterprise and sovereign AI lab. Raised at $6.8B, up from $5.5B in its 2024 round; hired former Meta AI research VP Joelle Pineau alongside the raise.","url":"https://cohere.com","sourceUrl":"https://cohere.com/blog/august-2025-funding-round"}
+{"id":"glean-2025-06","company":"Glean","category":"enterprise","stage":"series-f","amountM":150,"valuationB":7.2,"announcedDate":"2025-06-10","leadInvestors":["Wellington Management"],"notableInvestors":["Khosla Ventures","Altimeter","DST Global","Sequoia Capital"],"description":"Enterprise search and work AI assistant. Series F at $7.2B, up from $4.6B at the September 2024 Series E.","url":"https://www.glean.com","sourceUrl":"https://www.glean.com/blog/glean-series-f-announcement"}
+{"id":"cursor-2025-06","company":"Anysphere (Cursor)","category":"coding","stage":"series-c","amountM":900,"valuationB":9.9,"announcedDate":"2025-06-05","leadInvestors":["Thrive Capital"],"notableInvestors":["Andreessen Horowitz","Accel","DST Global"],"description":"AI-native code editor; passed $500M annualized revenue at the time of the round. SpaceX later agreed to acquire Anysphere in an all-stock deal valued at about $60B (signed June 16, 2026, closed August 14, 2026).","url":"https://cursor.com","sourceUrl":"https://cursor.com/blog/series-c"}
+{"id":"cartesia-2025-03","company":"Cartesia","category":"voice","stage":"series-a","amountM":64,"valuationB":null,"announcedDate":"2025-03-11","leadInvestors":["Kleiner Perkins"],"notableInvestors":["Index Ventures","Lightspeed","A* Capital"],"description":"Real-time voice AI (Sonic TTS models). Round launched Sonic 2.0 and brought total funding to $91M.","url":"https://cartesia.ai","sourceUrl":"https://www.cartesia.ai/blog/series-a"}
+{"id":"together-2025-q1","company":"Together AI","category":"inference","stage":"series-b","amountM":305,"valuationB":3.3,"announcedDate":"2025-02-20","leadInvestors":["General Catalyst","Prosperity7 Ventures"],"notableInvestors":["Salesforce Ventures","NVIDIA","Lux Capital"],"description":"Inference + fine-tuning hosting for open-weights models. Round funds GPU capacity expansion. Followed by an $800M Series C at $8.3B in July 2026.","url":"https://www.together.ai","sourceUrl":"https://www.together.ai/blog/series-b"}
+{"id":"harvey-2025-02","company":"Harvey","category":"enterprise","stage":"series-d","amountM":300,"valuationB":3,"announcedDate":"2025-02-12","leadInvestors":["Sequoia Capital"],"notableInvestors":["Coatue","Kleiner Perkins","GV (Google Ventures)","OpenAI Startup Fund","Conviction"],"description":"AI for legal work. Series D doubled the valuation from July 2024; RELX's REV (LexisNexis owner) joined as a strategic investor. Harvey went on to raise at $11B in March 2026 and $15.5B in September 2026.","url":"https://www.harvey.ai","sourceUrl":"https://www.harvey.ai/blog/harvey-raises-series-d"}
+{"id":"elevenlabs-2025-01","company":"ElevenLabs","category":"voice","stage":"series-c","amountM":180,"valuationB":3.3,"announcedDate":"2025-01-30","leadInvestors":["Andreessen Horowitz","ICONIQ Growth"],"notableInvestors":["NEA","Sequoia Capital","Salesforce Ventures"],"description":"Voice AI. Series C tripled the valuation in a year and brought total funding to $281M; strategic investors included Deutsche Telekom and NTT DOCOMO Ventures.","url":"https://elevenlabs.io","sourceUrl":"https://elevenlabs.io/blog/series-c"}
+{"id":"xai-2024-12","company":"xAI","category":"frontier-lab","stage":"series-c","amountM":6000,"valuationB":45,"announcedDate":"2024-12-23","leadInvestors":[],"notableInvestors":["Andreessen Horowitz","BlackRock","Fidelity","Kingdom Holdings","Sequoia Capital","Valor Equity Partners","NVIDIA","AMD"],"description":"Elon Musk's frontier lab; no lead was named. A Kingdom Holding filing put the valuation at $45B. xAI merged into SpaceX in February 2026.","url":"https://x.ai","sourceUrl":"https://x.ai/news/series-c"}
+{"id":"perplexity-2024-12","company":"Perplexity","category":"agent","stage":"growth","amountM":500,"valuationB":9,"announcedDate":"2024-12-18","leadInvestors":["IVP"],"notableInvestors":[],"description":"Answer engine. Round closed in December 2024 at $9B, tripling the June 2024 valuation; Perplexity declined to comment on it.","url":"https://www.perplexity.ai","sourceUrl":"https://www.bloomberg.com/news/articles/2024-12-18/ai-startup-perplexity-closes-funding-round-at-9-billion-value"}
+{"id":"crusoe-2024-12","company":"Crusoe","category":"infra","stage":"series-d","amountM":600,"valuationB":2.8,"announcedDate":"2024-12-12","leadInvestors":["Founders Fund"],"notableInvestors":["Fidelity","Mubadala","NVIDIA","Valor Equity Partners"],"description":"Vertically integrated AI data center and cloud operator. Followed by a $1.38B Series E at over $10B in October 2025 and a $3.9B Series F at $30.9B in September 2026 (separate row).","url":"https://crusoe.ai","sourceUrl":"https://www.crusoe.ai/resources/newsroom/crusoe-closes-series-d-funding"}
+{"id":"glean-2024-09","company":"Glean","category":"enterprise","stage":"series-e","amountM":260,"valuationB":4.6,"announcedDate":"2024-09-10","leadInvestors":["Altimeter","DST Global"],"notableInvestors":["Craft Ventures","Sapphire Ventures","SoftBank Vision Fund 2"],"description":"Enterprise search + AI assistant. Series E doubled the February 2024 valuation and predates the June 2025 Series F at $7.2B.","url":"https://www.glean.com","sourceUrl":"https://www.glean.com/press/glean-announces-over-260-million-series-e-and-next-generation-prompting-as-it-brings-work-ai-to-the-enterprise"}
+{"id":"sakana-2024-09","company":"Sakana AI","category":"frontier-lab","stage":"series-a","amountM":200,"valuationB":1.5,"announcedDate":"2024-09-04","leadInvestors":["New Enterprise Associates","Khosla Ventures","Lux Capital"],"notableInvestors":["NVIDIA","MUFG","SMBC","Mizuho"],"description":"Tokyo research lab (evolutionary model merging). Later raised a $135M Series B at $2.65B in November 2025, then took undisclosed strategic investments from Google (January 2026) and Mitsubishi Electric (March 2026).","url":"https://sakana.ai","sourceUrl":"https://sakana.ai/series-a/"}
+{"id":"groq-2024-08","company":"Groq","category":"inference","stage":"series-d","amountM":640,"valuationB":2.8,"announcedDate":"2024-08-05","leadInvestors":["BlackRock Private Equity Partners"],"notableInvestors":["Cisco Investments","Samsung Catalyst Fund","KDDI Open Innovation Fund","Neuberger Berman"],"description":"Custom LPU silicon for fast inference. Series D funded capacity expansion and more than doubled the 2021 valuation.","url":"https://groq.com","sourceUrl":"https://www.prnewswire.com/news-releases/groq-raises-640m-to-meet-soaring-demand-for-fast-ai-inference-302214097.html"}
+{"id":"openai-2024-10","company":"OpenAI","category":"frontier-lab","stage":"series-c","amountM":6600,"valuationB":157,"announcedDate":"2024-10-02","leadInvestors":["Thrive Capital"],"notableInvestors":["Microsoft","NVIDIA","Khosla Ventures","SoftBank"],"description":"$6.6B round at $157B post. Largest private round in tech history at the time.","url":"https://openai.com","sourceUrl":"https://openai.com/index/scale-the-benefits-of-ai/"}

2026-10-12/gpu-pricing.jsonl ADDED Viewed

	@@ -0,0 +1 @@

+ {"ok":true,"snapshot":{"capturedAt":"2026-10-12T16:15:56.282Z","providers":["runpod","lambda"],"offers":[{"provider":"runpod","gpu_raw":"MI350X","gpu_canonical":"OTHER","vram_gb":294,"on_demand_usd_hr":0.5,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"A100 PCIe","gpu_canonical":"A100-80GB","vram_gb":80,"on_demand_usd_hr":1.19,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"A100 SXM 40GB","gpu_canonical":"A100-40GB","vram_gb":40,"on_demand_usd_hr":1,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"A100 SXM","gpu_canonical":"A100-80GB","vram_gb":80,"on_demand_usd_hr":1.39,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"A40","gpu_canonical":"OTHER","vram_gb":48,"on_demand_usd_hr":0.35,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 3070","gpu_canonical":"OTHER","vram_gb":8,"on_demand_usd_hr":0.13,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 3080 Ti","gpu_canonical":"OTHER","vram_gb":12,"on_demand_usd_hr":0.18,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 3090","gpu_canonical":"RTX-3090","vram_gb":24,"on_demand_usd_hr":0.22,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 4070 Ti","gpu_canonical":"OTHER","vram_gb":12,"on_demand_usd_hr":0.19,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 4090","gpu_canonical":"RTX-4090","vram_gb":24,"on_demand_usd_hr":0.34,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 5080","gpu_canonical":"OTHER","vram_gb":16,"on_demand_usd_hr":0.39,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"H100 SXM","gpu_canonical":"H100","vram_gb":80,"on_demand_usd_hr":2.69,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"H100 NVL","gpu_canonical":"H100","vram_gb":80,"on_demand_usd_hr":2.59,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"H200 SXM","gpu_canonical":"H200","vram_gb":141,"on_demand_usd_hr":3.59,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"L40S","gpu_canonical":"L40S","vram_gb":48,"on_demand_usd_hr":0.79,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 4000 Ada SFF","gpu_canonical":"OTHER","vram_gb":20,"on_demand_usd_hr":0.18,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 5000 Ada","gpu_canonical":"OTHER","vram_gb":32,"on_demand_usd_hr":0.49,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX 6000 Ada","gpu_canonical":"RTX-6000-Ada","vram_gb":48,"on_demand_usd_hr":0.74,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX A4000","gpu_canonical":"A4000","vram_gb":16,"on_demand_usd_hr":0.17,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX PRO 4000","gpu_canonical":"OTHER","vram_gb":24,"on_demand_usd_hr":0.5,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX PRO 4500 SE","gpu_canonical":"OTHER","vram_gb":32,"on_demand_usd_hr":0.5,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX PRO 6000 MaxQ","gpu_canonical":"OTHER","vram_gb":96,"on_demand_usd_hr":1.64,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX PRO 6000","gpu_canonical":"OTHER","vram_gb":96,"on_demand_usd_hr":1.69,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"PRO 6000 MIG 24GB","gpu_canonical":"OTHER","vram_gb":24,"on_demand_usd_hr":0.5,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"PRO 6000 MIG 48GB","gpu_canonical":"OTHER","vram_gb":48,"on_demand_usd_hr":1,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"RTX PRO 6000 WK","gpu_canonical":"OTHER","vram_gb":96,"on_demand_usd_hr":1.69,"spot_usd_hr":null,"available_count":2,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"runpod","gpu_raw":"Tesla V100","gpu_canonical":"V100","vram_gb":32,"on_demand_usd_hr":0.19,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://runpod.io","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"B200 SXM6","gpu_canonical":"B200","vram_gb":192,"on_demand_usd_hr":6.69,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"H100 SXM","gpu_canonical":"H100","vram_gb":80,"on_demand_usd_hr":3.99,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"H100 PCIe","gpu_canonical":"H100","vram_gb":80,"on_demand_usd_hr":3.29,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"GH200","gpu_canonical":"H200","vram_gb":141,"on_demand_usd_hr":2.29,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"A100 SXM 80GB","gpu_canonical":"A100-80GB","vram_gb":80,"on_demand_usd_hr":2.79,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"A100 SXM 40GB","gpu_canonical":"A100-40GB","vram_gb":40,"on_demand_usd_hr":1.99,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"A100 PCIe 40GB","gpu_canonical":"A100-40GB","vram_gb":40,"on_demand_usd_hr":1.99,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"A10","gpu_canonical":"OTHER","vram_gb":24,"on_demand_usd_hr":1.29,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"A6000","gpu_canonical":"A6000","vram_gb":48,"on_demand_usd_hr":1.09,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"V100","gpu_canonical":"V100","vram_gb":32,"on_demand_usd_hr":0.79,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"},{"provider":"lambda","gpu_raw":"Quadro RTX 6000","gpu_canonical":"OTHER","vram_gb":24,"on_demand_usd_hr":0.69,"spot_usd_hr":null,"available_count":1,"region":null,"source_url":"https://lambda.ai/pricing","last_seen":"2026-10-12T16:15:56.621Z"}],"by_canonical":[{"canonical":"H200","vram_gb":141,"provider_count":2,"total_offers":2,"cheapest_on_demand":{"provider":"lambda","usd_hr":2.29,"gpu_raw":"GH200"},"cheapest_spot":null},{"canonical":"H100","vram_gb":80,"provider_count":2,"total_offers":4,"cheapest_on_demand":{"provider":"runpod","usd_hr":2.59,"gpu_raw":"H100 NVL"},"cheapest_spot":null},{"canonical":"B200","vram_gb":192,"provider_count":1,"total_offers":1,"cheapest_on_demand":{"provider":"lambda","usd_hr":6.69,"gpu_raw":"B200 SXM6"},"cheapest_spot":null},{"canonical":"A100-80GB","vram_gb":80,"provider_count":2,"total_offers":3,"cheapest_on_demand":{"provider":"runpod","usd_hr":1.19,"gpu_raw":"A100 PCIe"},"cheapest_spot":null},{"canonical":"A100-40GB","vram_gb":40,"provider_count":2,"total_offers":3,"cheapest_on_demand":{"provider":"runpod","usd_hr":1,"gpu_raw":"A100 SXM 40GB"},"cheapest_spot":null},{"canonical":"L40S","vram_gb":48,"provider_count":1,"total_offers":1,"cheapest_on_demand":{"provider":"runpod","usd_hr":0.79,"gpu_raw":"L40S"},"cheapest_spot":null},{"canonical":"RTX-6000-Ada","vram_gb":48,"provider_count":1,"total_offers":1,"cheapest_on_demand":{"provider":"runpod","usd_hr":0.74,"gpu_raw":"RTX 6000 Ada"},"cheapest_spot":null},{"canonical":"A6000","vram_gb":48,"provider_count":1,"total_offers":1,"cheapest_on_demand":{"provider":"lambda","usd_hr":1.09,"gpu_raw":"A6000"},"cheapest_spot":null},{"canonical":"A4000","vram_gb":16,"provider_count":1,"total_offers":1,"cheapest_on_demand":{"provider":"runpod","usd_hr":0.17,"gpu_raw":"RTX A4000"},"cheapest_spot":null},{"canonical":"RTX-4090","vram_gb":24,"provider_count":1,"total_offers":1,"cheapest_on_demand":{"provider":"runpod","usd_hr":0.34,"gpu_raw":"RTX 4090"},"cheapest_spot":null},{"canonical":"RTX-3090","vram_gb":24,"provider_count":1,"total_offers":1,"cheapest_on_demand":{"provider":"runpod","usd_hr":0.22,"gpu_raw":"RTX 3090"},"cheapest_spot":null},{"canonical":"V100","vram_gb":32,"provider_count":2,"total_offers":2,"cheapest_on_demand":{"provider":"runpod","usd_hr":0.19,"gpu_raw":"Tesla V100"},"cheapest_spot":null},{"canonical":"OTHER","vram_gb":null,"provider_count":2,"total_offers":17,"cheapest_on_demand":{"provider":"runpod","usd_hr":0.13,"gpu_raw":"RTX 3070"},"cheapest_spot":null}],"errors":[],"notes":["lambda pricing is a manual snapshot from 2026-05-06 (lambda.ai/pricing); bump the snapshot in code on redeploy when prices move"]}}

2026-10-12/harnesses.jsonl ADDED Viewed

	@@ -0,0 +1,6 @@

+{"id":"swe_bench_verified","name":"SWE-bench Verified","description":"500 human-validated GitHub issues across 12 Python repos. The harness must produce a patch that resolves the issue and passes the project's test suite.","maxScore":100,"unit":"% resolved","sourceUrl":"https://www.swebench.com/"}
+{"id":"terminal_bench","name":"Terminal-Bench","description":"Stanford and Anthropic benchmark of agentic terminal tasks. Each task gives the agent a goal and a sandboxed shell; success is measured by a deterministic post-condition.","maxScore":100,"unit":"% solved","sourceUrl":"https://www.tbench.ai/"}
+{"id":"terminal_bench_2_1","name":"Terminal-Bench 2.1","description":"The current revision of the Terminal-Bench shell-task suite, with repaired and rewritten tasks. Not comparable to the Terminal-Bench column: 2.1 is the harder revision, so the same harness and model pair scores lower on it.","maxScore":100,"unit":"% solved","sourceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.1"}
+{"id":"terminal_bench_4","name":"Terminal-Bench 4.0","description":"The current Terminal-Bench revision, hosted by Harbor and the Laude Institute. Scores sit far below the 2.1 column because the tasks are harder, so the two are separate measurements and must not be compared across columns.","maxScore":100,"unit":"% solved","sourceUrl":"https://www.tbench.ai/"}
+{"id":"aider_polyglot","name":"Aider Polyglot","description":"225 of the hardest Exercism coding exercises across C++, Go, Java, JavaScript, Python, and Rust. Measures whole-file edit-by-diff quality, not just code generation.","maxScore":100,"unit":"% pass2","sourceUrl":"https://aider.chat/docs/leaderboards/"}
+{"id":"swe_lancer","name":"SWE-Lancer","description":"OpenAI benchmark of paid Upwork engineering tasks ($1M+ in real bounties). Includes both diff-style fixes and longer feature work judged against the original buyer's acceptance criteria.","maxScore":100,"unit":"% earned","sourceUrl":"https://github.com/openai/SWELancer-Benchmark"}

2026-10-12/incidents.jsonl ADDED Viewed

	@@ -0,0 +1,224 @@

+{"id":"Claude API-1784064743634","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-14T21:32:23.634Z","resolvedAt":"2026-07-14T22:18:23.352Z","durationMinutes":46}
+{"id":"OpenAI API-1784073503550","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-14T23:58:23.550Z","resolvedAt":"2026-07-15T00:40:27.458Z","durationMinutes":42}
+{"id":"Claude API-1784124385340","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-15T14:06:25.340Z","resolvedAt":"2026-07-15T15:04:23.380Z","durationMinutes":58}
+{"id":"Claude API-1784191229045","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-16T08:40:29.045Z","resolvedAt":"2026-07-16T08:54:30.145Z","durationMinutes":14}
+{"id":"Claude API-1784192429045","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-16T09:00:29.045Z","resolvedAt":"2026-07-16T13:32:29.372Z","durationMinutes":272}
+{"id":"Luma-1784197597413","service":"Luma","provider":"Luma AI","severity":"minor","title":"Luma degraded performance","startedAt":"2026-07-16T10:26:37.413Z","resolvedAt":"2026-07-16T10:30:32.089Z","durationMinutes":4}
+{"id":"Claude API-1784227348663","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-16T18:42:28.663Z","resolvedAt":"2026-07-16T22:02:35.102Z","durationMinutes":200}
+{"id":"GitHub Copilot-1784236013047","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-07-16T21:06:53.047Z","resolvedAt":"2026-07-16T22:06:34.838Z","durationMinutes":60}
+{"id":"Claude API-1784270891284","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-17T06:48:11.284Z","resolvedAt":"2026-07-17T07:04:16.622Z","durationMinutes":16}
+{"id":"Claude API-1784272335381","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-17T07:12:15.381Z","resolvedAt":"2026-07-17T12:22:16.103Z","durationMinutes":310}
+{"id":"Claude API-1784300295668","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-17T14:58:15.668Z","resolvedAt":"2026-07-17T15:28:16.224Z","durationMinutes":30}
+{"id":"Claude API-1784312057156","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-17T18:14:17.156Z","resolvedAt":"2026-07-17T18:24:15.694Z","durationMinutes":10}
+{"id":"Claude API-1784313256611","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-17T18:34:16.611Z","resolvedAt":"2026-07-17T19:44:15.248Z","durationMinutes":70}
+{"id":"OpenAI API-1784473378664","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-19T15:02:58.664Z","resolvedAt":"2026-07-19T16:56:56.869Z","durationMinutes":114}
+{"id":"Claude API-1784531086501","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-20T07:04:46.501Z","resolvedAt":"2026-07-20T10:58:45.956Z","durationMinutes":234}
+{"id":"Claude API-1784553766308","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-20T13:22:46.308Z","resolvedAt":"2026-07-20T16:08:46.770Z","durationMinutes":166}
+{"id":"GitHub Copilot-1784563486919","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-07-20T16:04:46.919Z","resolvedAt":"2026-07-20T18:38:55.027Z","durationMinutes":154}
+{"id":"Claude API-1784617969768","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-21T07:12:49.768Z","resolvedAt":"2026-07-21T08:54:49.219Z","durationMinutes":102}
+{"id":"Claude API-1784628769395","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-21T10:12:49.395Z","resolvedAt":"2026-07-21T11:10:49.571Z","durationMinutes":58}
+{"id":"Claude API-1784632849381","service":"Claude API","provider":"Anthropic","severity":"major","title":"Claude API outage","startedAt":"2026-07-21T11:20:49.381Z","resolvedAt":"2026-07-21T11:48:49.215Z","durationMinutes":28}
+{"id":"Claude API-1784639210600","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-21T13:06:50.600Z","resolvedAt":"2026-07-21T13:24:49.714Z","durationMinutes":18}
+{"id":"Claude API-1784645209903","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-21T14:46:49.903Z","resolvedAt":"2026-07-21T17:08:15.780Z","durationMinutes":141}
+{"id":"Claude API-1784655736648","service":"Claude API","provider":"Anthropic","severity":"major","title":"Claude API outage","startedAt":"2026-07-21T17:42:16.648Z","resolvedAt":"2026-07-21T18:04:15.557Z","durationMinutes":22}
+{"id":"OpenAI API-1784663535801","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-21T19:52:15.801Z","resolvedAt":"2026-07-21T20:28:15.720Z","durationMinutes":36}
+{"id":"Claude API-1784709403354","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-22T08:36:43.354Z","resolvedAt":"2026-07-22T08:46:43.833Z","durationMinutes":10}
+{"id":"Replicate-1784721163339","service":"Replicate","provider":"Replicate","severity":"minor","title":"Replicate degraded performance","startedAt":"2026-07-22T11:52:43.339Z","resolvedAt":"2026-07-23T03:06:08.969Z","durationMinutes":913}
+{"id":"Claude API-1784727524180","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-22T13:38:44.180Z","resolvedAt":"2026-07-22T15:08:47.316Z","durationMinutes":90}
+{"id":"Claude API-1784733884275","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-22T15:24:44.275Z","resolvedAt":"2026-07-22T16:12:44.427Z","durationMinutes":48}
+{"id":"Claude API-1784741338530","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-22T17:28:58.530Z","resolvedAt":"2026-07-22T17:46:58.510Z","durationMinutes":18}
+{"id":"Claude API-1784764449142","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-22T23:54:09.142Z","resolvedAt":"2026-07-23T00:12:09.114Z","durationMinutes":18}
+{"id":"OpenAI API-1784821013672","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-23T15:36:53.672Z","resolvedAt":"2026-07-23T17:54:52.757Z","durationMinutes":138}
+{"id":"Perplexity-1784829052840","service":"Perplexity","provider":"Perplexity AI","severity":"minor","title":"Perplexity degraded performance","startedAt":"2026-07-23T17:50:52.840Z","resolvedAt":"2026-07-23T18:14:52.743Z","durationMinutes":24}
+{"id":"OpenAI API-1784829412778","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-23T17:56:52.778Z","resolvedAt":"2026-07-23T18:50:52.995Z","durationMinutes":54}
+{"id":"OpenAI API-1784834212905","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-23T19:16:52.905Z","resolvedAt":"2026-07-23T20:24:08.218Z","durationMinutes":67}
+{"id":"OpenRouter-1784840405945","service":"OpenRouter","provider":"OpenRouter","severity":"major","title":"OpenRouter outage","startedAt":"2026-07-23T21:00:05.945Z","resolvedAt":"2026-07-23T21:08:06.327Z","durationMinutes":8}
+{"id":"OpenAI API-1784851092127","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-23T23:58:12.127Z","resolvedAt":"2026-07-24T00:10:12.088Z","durationMinutes":12}
+{"id":"Claude API-1784884332683","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-24T09:12:12.683Z","resolvedAt":"2026-07-24T09:36:11.473Z","durationMinutes":24}
+{"id":"OpenAI API-1784887572080","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-24T10:06:12.080Z","resolvedAt":"2026-07-24T11:36:11.506Z","durationMinutes":90}
+{"id":"GitHub Copilot-1784910492148","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-07-24T16:28:12.148Z","resolvedAt":"2026-07-24T17:18:12.279Z","durationMinutes":50}
+{"id":"OpenAI API-1784971137974","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-25T09:18:57.974Z","resolvedAt":"2026-07-25T11:08:58.412Z","durationMinutes":110}
+{"id":"GitHub Copilot-1784972698580","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-07-25T09:44:58.580Z","resolvedAt":"2026-07-25T10:06:57.982Z","durationMinutes":22}
+{"id":"OpenAI API-1784979418086","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-07-25T11:36:58.086Z","resolvedAt":"2026-07-25T11:58:58.525Z","durationMinutes":22}
+{"id":"Claude API-1785004858194","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-25T18:40:58.194Z","resolvedAt":"2026-07-25T19:44:58.736Z","durationMinutes":64}
+{"id":"Claude API-1785015298360","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-25T21:34:58.360Z","resolvedAt":"2026-07-25T22:10:58.440Z","durationMinutes":36}
+{"id":"Claude API-1785057484562","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-26T09:18:04.562Z","resolvedAt":"2026-07-26T10:38:04.055Z","durationMinutes":80}
+{"id":"Claude API-1785140284378","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-27T08:18:04.378Z","resolvedAt":"2026-07-27T09:06:05.753Z","durationMinutes":48}
+{"id":"Claude API-1785151684324","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-27T11:28:04.324Z","resolvedAt":"2026-07-27T12:32:05.659Z","durationMinutes":64}
+{"id":"Claude API-1785159605753","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-27T13:40:05.753Z","resolvedAt":"2026-07-27T14:38:04.564Z","durationMinutes":58}
+{"id":"Claude API-1785354643958","service":"Claude API","provider":"Anthropic","severity":"major","title":"Claude API outage","startedAt":"2026-07-29T19:50:43.958Z","resolvedAt":"2026-07-29T22:22:43.846Z","durationMinutes":152}
+{"id":"GitHub Copilot-1785355732316","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-07-29T20:08:52.316Z","resolvedAt":"2026-07-29T21:52:44.090Z","durationMinutes":104}
+{"id":"Claude API-1785391130598","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-30T05:58:50.598Z","resolvedAt":"2026-07-30T10:04:43.853Z","durationMinutes":246}
+{"id":"Cursor-1785392205560","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-07-30T06:16:45.560Z","resolvedAt":"2026-07-30T06:34:43.408Z","durationMinutes":18}
+{"id":"Cursor-1785401924213","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-07-30T08:58:44.213Z","resolvedAt":"2026-07-30T09:50:44.032Z","durationMinutes":52}
+{"id":"GitHub Copilot-1785402524442","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-07-30T09:08:44.442Z","resolvedAt":"2026-07-30T10:14:43.344Z","durationMinutes":66}
+{"id":"Claude API-1785419084076","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-30T13:44:44.076Z","resolvedAt":"2026-07-30T14:26:44.238Z","durationMinutes":42}
+{"id":"ElevenLabs-1785467448832","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-07-31T03:10:48.832Z","resolvedAt":"2026-07-31T05:30:46.571Z","durationMinutes":140}
+{"id":"Claude API-1785478805435","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-07-31T06:20:05.435Z","resolvedAt":"2026-07-31T07:06:06.217Z","durationMinutes":46}
+{"id":"ElevenLabs-1785478805435","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-07-31T06:20:05.435Z","resolvedAt":"2026-07-31T19:17:00.896Z","durationMinutes":777}
+{"id":"Cursor-1785527460054","service":"Cursor","provider":"Cursor","severity":"major","title":"Cursor outage","startedAt":"2026-07-31T19:51:00.054Z","resolvedAt":"2026-07-31T21:11:00.237Z","durationMinutes":80}
+{"id":"Replicate-1785533220193","service":"Replicate","provider":"Replicate","severity":"minor","title":"Replicate degraded performance","startedAt":"2026-07-31T21:27:00.193Z","resolvedAt":"2026-07-31T22:29:00.406Z","durationMinutes":62}
+{"id":"GitHub Copilot-1785583114402","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-08-01T11:18:34.402Z","resolvedAt":"2026-08-01T12:32:34.415Z","durationMinutes":74}
+{"id":"GitHub Copilot-1785607474384","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-08-01T18:04:34.384Z","resolvedAt":"2026-08-01T18:22:34.849Z","durationMinutes":18}
+{"id":"GitHub Copilot-1785750881403","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-08-03T09:54:41.403Z","resolvedAt":"2026-08-03T11:20:40.935Z","durationMinutes":86}
+{"id":"Claude API-1785761681630","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-03T12:54:41.630Z","resolvedAt":"2026-08-03T13:32:42.687Z","durationMinutes":38}
+{"id":"Claude API-1785770082391","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-03T15:14:42.391Z","resolvedAt":"2026-08-03T15:30:41.549Z","durationMinutes":16}
+{"id":"Claude API-1785824321631","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-04T06:18:41.631Z","resolvedAt":"2026-08-04T06:28:42.202Z","durationMinutes":10}
+{"id":"Claude API-1785842921688","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-04T11:28:41.688Z","resolvedAt":"2026-08-04T11:40:41.130Z","durationMinutes":12}
+{"id":"Claude API-1785876641258","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-04T20:50:41.258Z","resolvedAt":"2026-08-04T21:06:40.981Z","durationMinutes":16}
+{"id":"Cursor-1785886601529","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-04T23:36:41.529Z","resolvedAt":"2026-08-05T00:02:43.750Z","durationMinutes":26}
+{"id":"Claude API-1785913715006","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-05T07:08:35.006Z","resolvedAt":"2026-08-05T13:10:34.365Z","durationMinutes":362}
+{"id":"Cursor-1785917077369","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-05T08:04:37.369Z","resolvedAt":"2026-08-05T08:08:33.868Z","durationMinutes":4}
+{"id":"Cursor-1785917433689","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-05T08:10:33.689Z","resolvedAt":"2026-08-05T14:36:34.979Z","durationMinutes":386}
+{"id":"Claude API-1785937954338","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-05T13:52:34.338Z","resolvedAt":"2026-08-05T14:16:34.854Z","durationMinutes":24}
+{"id":"OpenAI API-1785946594679","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-08-05T16:16:34.679Z","resolvedAt":"2026-08-05T17:52:34.418Z","durationMinutes":96}
+{"id":"Cursor-1785947314783","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-05T16:28:34.783Z","resolvedAt":"2026-08-05T18:48:34.444Z","durationMinutes":140}
+{"id":"Perplexity-1786033577735","service":"Perplexity","provider":"Perplexity AI","severity":"minor","title":"Perplexity degraded performance","startedAt":"2026-08-06T16:26:17.735Z","resolvedAt":"2026-08-06T18:10:19.675Z","durationMinutes":104}
+{"id":"Cursor-1786046429887","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-06T20:00:29.887Z","resolvedAt":"2026-08-06T20:28:19.738Z","durationMinutes":28}
+{"id":"Cursor-1786118056672","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-07T15:54:16.672Z","resolvedAt":"2026-08-07T19:50:22.374Z","durationMinutes":236}
+{"id":"Replicate-1786395968298","service":"Replicate","provider":"Replicate","severity":"minor","title":"Replicate degraded performance","startedAt":"2026-08-10T21:06:08.298Z","resolvedAt":"2026-08-10T21:46:08.703Z","durationMinutes":40}
+{"id":"Cursor-1786469048549","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-11T17:24:08.549Z","resolvedAt":"2026-08-11T23:40:08.132Z","durationMinutes":376}
+{"id":"Cursor-1786498828392","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-12T01:40:28.392Z","resolvedAt":"2026-08-12T04:12:27.857Z","durationMinutes":152}
+{"id":"Cursor-1786511428117","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-12T05:10:28.117Z","resolvedAt":"2026-08-12T05:30:29.800Z","durationMinutes":20}
+{"id":"Claude API-1786542750112","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-12T13:52:30.112Z","resolvedAt":"2026-08-12T18:08:28.143Z","durationMinutes":256}
+{"id":"Cursor-1786568188339","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-12T20:56:28.339Z","resolvedAt":"2026-08-12T21:14:28.440Z","durationMinutes":18}
+{"id":"Perplexity-1786584268772","service":"Perplexity","provider":"Perplexity AI","severity":"minor","title":"Perplexity degraded performance","startedAt":"2026-08-13T01:24:28.772Z","resolvedAt":"2026-08-13T01:50:28.196Z","durationMinutes":26}
+{"id":"Runway-1786591708447","service":"Runway","provider":"Runway","severity":"minor","title":"Runway degraded performance","startedAt":"2026-08-13T03:28:28.447Z","resolvedAt":"2026-08-13T03:40:29.564Z","durationMinutes":12}
+{"id":"Claude API-1786631701020","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-13T14:35:01.020Z","resolvedAt":"2026-08-13T18:00:16.474Z","durationMinutes":205}
+{"id":"GitHub Copilot-1786632299595","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-08-13T14:44:59.595Z","resolvedAt":"2026-08-13T15:49:00.400Z","durationMinutes":64}
+{"id":"Cursor-1786632540674","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-13T14:49:00.674Z","resolvedAt":"2026-08-13T18:50:16.419Z","durationMinutes":241}
+{"id":"Perplexity-1786708350705","service":"Perplexity","provider":"Perplexity AI","severity":"minor","title":"Perplexity degraded performance","startedAt":"2026-08-14T11:52:30.705Z","resolvedAt":"2026-08-14T11:58:28.063Z","durationMinutes":6}
+{"id":"Claude API-1786739548128","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-14T20:32:28.128Z","resolvedAt":"2026-08-14T22:42:28.549Z","durationMinutes":130}
+{"id":"Claude API-1786751907767","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-14T23:58:27.767Z","resolvedAt":"2026-08-15T00:22:27.925Z","durationMinutes":24}
+{"id":"OpenAI API-1786757308586","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-08-15T01:28:28.586Z","resolvedAt":"2026-08-15T08:50:27.717Z","durationMinutes":442}
+{"id":"Claude API-1786917609758","service":"Claude API","provider":"Anthropic","severity":"major","title":"Claude API outage","startedAt":"2026-08-16T22:00:09.758Z","resolvedAt":"2026-08-16T22:36:06.303Z","durationMinutes":36}
+{"id":"Cursor-1786959177645","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-17T09:32:57.645Z","resolvedAt":"2026-08-17T13:40:57.566Z","durationMinutes":248}
+{"id":"Claude API-1786975137702","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-17T13:58:57.702Z","resolvedAt":"2026-08-17T14:40:57.492Z","durationMinutes":42}
+{"id":"GitHub Copilot-1786977177699","service":"GitHub Copilot","provider":"GitHub","severity":"major","title":"GitHub Copilot outage","startedAt":"2026-08-17T14:32:57.699Z","resolvedAt":"2026-08-17T21:16:26.868Z","durationMinutes":403}
+{"id":"Cursor-1786977297614","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-17T14:34:57.614Z","resolvedAt":"2026-08-17T20:40:26.464Z","durationMinutes":365}
+{"id":"Cursor-1787003538722","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-17T21:52:18.722Z","resolvedAt":"2026-08-17T22:58:19.401Z","durationMinutes":66}
+{"id":"Cursor-1787039059431","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-18T07:44:19.431Z","resolvedAt":"2026-08-18T08:52:28.142Z","durationMinutes":68}
+{"id":"Cursor-1787049500493","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-18T10:38:20.493Z","resolvedAt":"2026-08-18T11:46:20.497Z","durationMinutes":68}
+{"id":"ElevenLabs-1787056818870","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-08-18T12:40:18.870Z","resolvedAt":"2026-08-18T12:42:19.860Z","durationMinutes":2}
+{"id":"Claude API-1787070139266","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-18T16:22:19.266Z","resolvedAt":"2026-08-18T18:28:19.619Z","durationMinutes":126}
+{"id":"Cursor-1787078899601","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-18T18:48:19.601Z","resolvedAt":"2026-08-19T02:44:21.220Z","durationMinutes":476}
+{"id":"Claude API-1787132660474","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-19T09:44:20.474Z","resolvedAt":"2026-08-19T10:34:19.128Z","durationMinutes":50}
+{"id":"Cursor-1787144061526","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-19T12:54:21.526Z","resolvedAt":"2026-08-19T15:56:20.170Z","durationMinutes":182}
+{"id":"Runway-1787158699267","service":"Runway","provider":"Runway","severity":"minor","title":"Runway degraded performance","startedAt":"2026-08-19T16:58:19.267Z","resolvedAt":"2026-08-19T18:04:19.026Z","durationMinutes":66}
+{"id":"Cursor-1787181139156","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-19T23:12:19.156Z","resolvedAt":"2026-08-19T23:16:19.966Z","durationMinutes":4}
+{"id":"Cursor-1787181499044","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-19T23:18:19.044Z","resolvedAt":"2026-08-20T01:52:19.360Z","durationMinutes":154}
+{"id":"Claude API-1787250874928","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-20T18:34:34.928Z","resolvedAt":"2026-08-20T19:02:35.397Z","durationMinutes":28}
+{"id":"Cursor-1787304126832","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-21T09:22:06.832Z","resolvedAt":"2026-08-21T10:10:07.485Z","durationMinutes":48}
+{"id":"Claude API-1787548132208","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-24T05:08:52.208Z","resolvedAt":"2026-08-24T08:32:52.363Z","durationMinutes":204}
+{"id":"Cursor-1787550052489","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-24T05:40:52.489Z","resolvedAt":"2026-08-24T08:52:33.007Z","durationMinutes":192}
+{"id":"GitHub Copilot-1787555573104","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-08-24T07:12:53.104Z","resolvedAt":"2026-08-24T08:00:55.286Z","durationMinutes":48}
+{"id":"ElevenLabs-1787561433535","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-08-24T08:50:33.535Z","resolvedAt":"2026-08-24T18:44:33.549Z","durationMinutes":594}
+{"id":"Cursor-1787587593485","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-24T16:06:33.485Z","resolvedAt":"2026-08-24T16:34:41.935Z","durationMinutes":28}
+{"id":"Claude API-1787591673868","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-24T17:14:33.868Z","resolvedAt":"2026-08-24T17:18:33.478Z","durationMinutes":4}
+{"id":"Claude API-1787602353170","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-24T20:12:33.170Z","resolvedAt":"2026-08-24T20:28:33.720Z","durationMinutes":16}
+{"id":"Cursor-1787669674135","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-25T14:54:34.135Z","resolvedAt":"2026-08-25T17:38:33.442Z","durationMinutes":164}
+{"id":"Cursor-1787690188144","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-25T20:36:28.144Z","resolvedAt":"2026-08-25T21:30:36.732Z","durationMinutes":54}
+{"id":"ElevenLabs-1787764616886","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-08-26T17:16:56.886Z","resolvedAt":"2026-08-26T17:54:56.293Z","durationMinutes":38}
+{"id":"Replicate-1787779136214","service":"Replicate","provider":"Replicate","severity":"minor","title":"Replicate degraded performance","startedAt":"2026-08-26T21:18:56.214Z","resolvedAt":"2026-08-27T00:26:56.078Z","durationMinutes":188}
+{"id":"GitHub Copilot-1787825173843","service":"GitHub Copilot","provider":"GitHub","severity":"major","title":"GitHub Copilot outage","startedAt":"2026-08-27T10:06:13.843Z","resolvedAt":"2026-08-27T12:14:13.669Z","durationMinutes":128}
+{"id":"Cursor-1787840653285","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-08-27T14:24:13.285Z","resolvedAt":"2026-08-27T15:06:13.375Z","durationMinutes":42}
+{"id":"Cursor-1787859133634","service":"Cursor","provider":"Cursor","severity":"major","title":"Cursor outage","startedAt":"2026-08-27T19:32:13.634Z","resolvedAt":"2026-08-27T19:40:13.878Z","durationMinutes":8}
+{"id":"Runway-1787933415538","service":"Runway","provider":"Runway","severity":"minor","title":"Runway degraded performance","startedAt":"2026-08-28T16:10:15.538Z","resolvedAt":"2026-08-28T21:18:49.412Z","durationMinutes":309}
+{"id":"OpenRouter-1787946735325","service":"OpenRouter","provider":"OpenRouter","severity":"major","title":"OpenRouter outage","startedAt":"2026-08-28T19:52:15.325Z","resolvedAt":"2026-08-28T22:34:49.510Z","durationMinutes":163}
+{"id":"ElevenLabs-1788163856247","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-08-31T08:10:56.247Z","resolvedAt":"2026-08-31T08:14:56.624Z","durationMinutes":4}
+{"id":"GitHub Copilot-1788167816666","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-08-31T09:16:56.666Z","resolvedAt":"2026-08-31T09:52:56.642Z","durationMinutes":36}
+{"id":"Claude API-1788197096496","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-31T17:24:56.496Z","resolvedAt":"2026-08-31T17:34:56.474Z","durationMinutes":10}
+{"id":"Claude API-1788197816314","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-31T17:36:56.314Z","resolvedAt":"2026-08-31T17:40:56.820Z","durationMinutes":4}
+{"id":"Claude API-1788204167822","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-08-31T19:22:47.822Z","resolvedAt":"2026-08-31T20:38:47.005Z","durationMinutes":76}
+{"id":"OpenAI API-1788215326974","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-08-31T22:28:46.974Z","resolvedAt":"2026-09-01T19:06:44.261Z","durationMinutes":1238}
+{"id":"Perplexity-1788220727100","service":"Perplexity","provider":"Perplexity AI","severity":"minor","title":"Perplexity degraded performance","startedAt":"2026-08-31T23:58:47.100Z","resolvedAt":"2026-09-01T00:08:46.724Z","durationMinutes":10}
+{"id":"Cursor-1788224564444","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-01T01:02:44.444Z","resolvedAt":"2026-09-01T01:14:44.078Z","durationMinutes":12}
+{"id":"Cohere-1788277364264","service":"Cohere","provider":"Cohere","severity":"minor","title":"Cohere degraded performance","startedAt":"2026-09-01T15:42:44.264Z","resolvedAt":"2026-09-01T19:48:44.125Z","durationMinutes":246}
+{"id":"Claude API-1788278804066","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-01T16:06:44.066Z","resolvedAt":"2026-09-01T16:24:44.174Z","durationMinutes":18}
+{"id":"Cursor-1788280964236","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-01T16:42:44.236Z","resolvedAt":"2026-09-01T18:42:44.027Z","durationMinutes":120}
+{"id":"Claude API-1788282404289","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-01T17:06:44.289Z","resolvedAt":"2026-09-01T18:08:44.119Z","durationMinutes":62}
+{"id":"Cursor-1788310124370","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-02T00:48:44.370Z","resolvedAt":"2026-09-02T04:08:15.454Z","durationMinutes":200}
+{"id":"Cursor-1788360754434","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-02T14:52:34.434Z","resolvedAt":"2026-09-02T16:12:34.220Z","durationMinutes":80}
+{"id":"Claude API-1788384172022","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-02T21:22:52.022Z","resolvedAt":"2026-09-02T21:46:52.243Z","durationMinutes":24}
+{"id":"OpenAI API-1788393891532","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-09-03T00:04:51.532Z","resolvedAt":"2026-09-03T00:12:52.546Z","durationMinutes":8}
+{"id":"Claude API-1788439115838","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-03T12:38:35.838Z","resolvedAt":"2026-09-03T12:48:36.884Z","durationMinutes":10}
+{"id":"Claude API-1788442004699","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-03T13:26:44.699Z","resolvedAt":"2026-09-03T16:24:47.326Z","durationMinutes":178}
+{"id":"Cursor-1788442954925","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-03T13:42:34.925Z","resolvedAt":"2026-09-03T17:08:48.256Z","durationMinutes":206}
+{"id":"GitHub Copilot-1788445115728","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-09-03T14:18:35.728Z","resolvedAt":"2026-09-03T17:12:46.986Z","durationMinutes":174}
+{"id":"OpenAI API-1788447516781","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-09-03T14:58:36.781Z","resolvedAt":"2026-09-03T16:56:46.418Z","durationMinutes":118}
+{"id":"Cursor-1788457478253","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-03T17:44:38.253Z","resolvedAt":"2026-09-03T19:22:38.916Z","durationMinutes":98}
+{"id":"Perplexity-1788467679651","service":"Perplexity","provider":"Perplexity AI","severity":"minor","title":"Perplexity degraded performance","startedAt":"2026-09-03T20:34:39.651Z","resolvedAt":"2026-09-03T21:14:39.005Z","durationMinutes":40}
+{"id":"Cursor-1788479078779","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-03T23:44:38.779Z","resolvedAt":"2026-09-04T00:44:38.778Z","durationMinutes":60}
+{"id":"ElevenLabs-1788522140443","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-09-04T11:42:20.443Z","resolvedAt":"2026-09-04T12:12:21.894Z","durationMinutes":30}
+{"id":"ElevenLabs-1788530183157","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-09-04T13:56:23.157Z","resolvedAt":"2026-09-04T21:02:20.472Z","durationMinutes":426}
+{"id":"Perplexity-1788587421063","service":"Perplexity","provider":"Perplexity AI","severity":"minor","title":"Perplexity degraded performance","startedAt":"2026-09-05T05:50:21.063Z","resolvedAt":"2026-09-05T07:10:26.527Z","durationMinutes":80}
+{"id":"Cursor-1788813269608","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-07T20:34:29.608Z","resolvedAt":"2026-09-07T21:38:29.511Z","durationMinutes":64}
+{"id":"OpenAI API-1788878189532","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-09-08T14:36:29.532Z","resolvedAt":"2026-09-08T22:00:30.221Z","durationMinutes":444}
+{"id":"ElevenLabs-1788957749303","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-09-09T12:42:29.303Z","resolvedAt":"2026-09-10T20:44:55.230Z","durationMinutes":1922}
+{"id":"Cursor-1789060135886","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-10T17:08:55.886Z","resolvedAt":"2026-09-10T17:32:57.923Z","durationMinutes":24}
+{"id":"Claude API-1789076695192","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-10T21:44:55.192Z","resolvedAt":"2026-09-10T23:24:55.166Z","durationMinutes":100}
+{"id":"Cursor-1789125062554","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-11T11:11:02.554Z","resolvedAt":"2026-09-11T11:36:54.723Z","durationMinutes":26}
+{"id":"Claude API-1789136453070","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-11T14:20:53.070Z","resolvedAt":"2026-09-11T14:30:54.927Z","durationMinutes":10}
+{"id":"Cursor-1789373804475","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-14T08:16:44.475Z","resolvedAt":"2026-09-14T09:08:48.543Z","durationMinutes":52}
+{"id":"Stability AI-1789403678761","service":"Stability AI","provider":"Stability AI","severity":"minor","title":"Stability AI degraded performance","startedAt":"2026-09-14T16:34:38.761Z","resolvedAt":"2026-09-14T17:56:32.616Z","durationMinutes":82}
+{"id":"Claude API-1789465831425","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-15T09:50:31.425Z","resolvedAt":"2026-09-15T11:16:32.457Z","durationMinutes":86}
+{"id":"GitHub Copilot-1789467752084","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-09-15T10:22:32.084Z","resolvedAt":"2026-09-15T11:18:31.234Z","durationMinutes":56}
+{"id":"Cursor-1789507711500","service":"Cursor","provider":"Cursor","severity":"major","title":"Cursor outage","startedAt":"2026-09-15T21:28:31.500Z","resolvedAt":"2026-09-15T22:00:32.409Z","durationMinutes":32}
+{"id":"OpenAI API-1789516352549","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-09-15T23:52:32.549Z","resolvedAt":"2026-09-16T00:32:32.011Z","durationMinutes":40}
+{"id":"GitHub Copilot-1789543352107","service":"GitHub Copilot","provider":"GitHub","severity":"minor","title":"GitHub Copilot degraded performance","startedAt":"2026-09-16T07:22:32.107Z","resolvedAt":"2026-09-16T11:46:32.103Z","durationMinutes":264}
+{"id":"Cursor-1789545872536","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-16T08:04:32.536Z","resolvedAt":"2026-09-16T08:44:33.053Z","durationMinutes":40}
+{"id":"Cursor-1789568912099","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-16T14:28:32.099Z","resolvedAt":"2026-09-16T20:58:31.713Z","durationMinutes":390}
+{"id":"Cursor-1789648342280","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-17T12:32:22.280Z","resolvedAt":"2026-09-17T14:36:29.733Z","durationMinutes":124}
+{"id":"Cursor-1789658546610","service":"Cursor","provider":"Cursor","severity":"major","title":"Cursor outage","startedAt":"2026-09-17T15:22:26.610Z","resolvedAt":"2026-09-17T20:04:25.457Z","durationMinutes":282}
+{"id":"OpenAI API-1789680640390","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-09-17T21:30:40.390Z","resolvedAt":"2026-09-17T21:34:23.390Z","durationMinutes":4}
+{"id":"Cursor-1789731537771","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-18T11:38:57.771Z","resolvedAt":"2026-09-18T14:34:51.158Z","durationMinutes":176}
+{"id":"Cursor-1789752083773","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-18T17:21:23.773Z","resolvedAt":"2026-09-18T23:14:22.192Z","durationMinutes":353}
+{"id":"ElevenLabs-1789778061796","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-09-19T00:34:21.796Z","resolvedAt":"2026-09-19T00:38:23.295Z","durationMinutes":4}
+{"id":"ElevenLabs-1789779024084","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-09-19T00:50:24.084Z","resolvedAt":"2026-09-19T00:58:22.135Z","durationMinutes":8}
+{"id":"Cursor-1789972830851","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-21T06:40:30.851Z","resolvedAt":"2026-09-21T07:02:21.361Z","durationMinutes":22}
+{"id":"Claude API-1790038709169","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-22T00:58:29.169Z","resolvedAt":"2026-09-22T02:12:27.334Z","durationMinutes":74}
+{"id":"Cursor-1790039310115","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-22T01:08:30.115Z","resolvedAt":"2026-09-22T01:44:19.602Z","durationMinutes":36}
+{"id":"Cursor-1790082181866","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-22T13:03:01.866Z","resolvedAt":"2026-09-22T13:46:45.606Z","durationMinutes":44}
+{"id":"Cursor-1790183795094","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-23T17:16:35.094Z","resolvedAt":"2026-09-23T18:44:24.230Z","durationMinutes":88}
+{"id":"Cursor-1790189666098","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-23T18:54:26.098Z","resolvedAt":"2026-09-23T18:56:27.688Z","durationMinutes":2}
+{"id":"Cursor-1790195302793","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-23T20:28:22.793Z","resolvedAt":"2026-09-23T20:38:25.793Z","durationMinutes":10}
+{"id":"Cursor-1790313537652","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-25T05:18:57.652Z","resolvedAt":"2026-09-25T06:52:33.707Z","durationMinutes":94}
+{"id":"Cursor-1790380143517","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-25T23:49:03.517Z","resolvedAt":"2026-09-26T00:35:03.091Z","durationMinutes":46}
+{"id":"ElevenLabs-1790610178877","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-09-28T15:42:58.877Z","resolvedAt":"2026-09-28T16:07:07.759Z","durationMinutes":24}
+{"id":"Cursor-1790627068077","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-28T20:24:28.077Z","resolvedAt":"2026-09-28T20:36:59.008Z","durationMinutes":13}
+{"id":"Azure OpenAI-1790689252549","service":"Azure OpenAI","provider":"Microsoft Azure","severity":"minor","title":"Azure OpenAI degraded performance","startedAt":"2026-09-29T13:40:52.549Z","resolvedAt":"2026-09-29T18:56:28.005Z","durationMinutes":316}
+{"id":"Claude API-1790691780867","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-09-29T14:23:00.867Z","resolvedAt":"2026-09-29T16:29:00.850Z","durationMinutes":126}
+{"id":"Cursor-1790692020622","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-29T14:27:00.622Z","resolvedAt":"2026-09-29T18:07:00.637Z","durationMinutes":220}
+{"id":"OpenAI API-1790704492101","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-09-29T17:54:52.101Z","resolvedAt":"2026-09-29T23:14:27.776Z","durationMinutes":320}
+{"id":"ElevenLabs-1790707828468","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-09-29T18:50:28.468Z","resolvedAt":"2026-09-29T20:24:27.512Z","durationMinutes":94}
+{"id":"Cursor-1790711067747","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-09-29T19:44:27.747Z","resolvedAt":"2026-09-29T20:20:28.602Z","durationMinutes":36}
+{"id":"OpenAI API-1790787988409","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-09-30T17:06:28.409Z","resolvedAt":"2026-09-30T20:36:15.806Z","durationMinutes":210}
+{"id":"Claude API-1790871744340","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-10-01T16:22:24.340Z","resolvedAt":"2026-10-01T16:24:24.646Z","durationMinutes":2}
+{"id":"Claude API-1790871975850","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-10-01T16:26:15.850Z","resolvedAt":"2026-10-01T18:24:16.091Z","durationMinutes":118}
+{"id":"Cursor-1791010377256","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-03T06:52:57.256Z","resolvedAt":"2026-10-03T07:28:58.942Z","durationMinutes":36}
+{"id":"AWS Bedrock-1791133021371","service":"AWS Bedrock","provider":"AWS","severity":"minor","title":"AWS Bedrock degraded performance","startedAt":"2026-10-04T16:57:01.371Z","resolvedAt":"2026-10-05T18:58:04.594Z","durationMinutes":1561}
+{"id":"OpenAI API-1791183842093","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-10-05T07:04:02.093Z","resolvedAt":"2026-10-06T08:30:33.644Z","durationMinutes":1527}
+{"id":"Claude API-1791205581266","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-10-05T13:06:21.266Z","resolvedAt":"2026-10-05T13:22:04.879Z","durationMinutes":16}
+{"id":"ElevenLabs-1791277833564","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-10-06T09:10:33.564Z","resolvedAt":"2026-10-06T12:52:33.860Z","durationMinutes":222}
+{"id":"Claude API-1791289596701","service":"Claude API","provider":"Anthropic","severity":"minor","title":"Claude API degraded performance","startedAt":"2026-10-06T12:26:36.701Z","resolvedAt":"2026-10-06T12:44:42.913Z","durationMinutes":18}
+{"id":"Cursor-1791290496319","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-06T12:41:36.319Z","resolvedAt":"2026-10-06T16:04:35.847Z","durationMinutes":203}
+{"id":"OpenAI API-1791301627250","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-10-06T15:47:07.250Z","resolvedAt":"2026-10-06T20:12:33.463Z","durationMinutes":265}
+{"id":"Cursor-1791309034258","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-06T17:50:34.258Z","resolvedAt":"2026-10-06T18:58:33.808Z","durationMinutes":68}
+{"id":"Cursor-1791314914444","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-06T19:28:34.444Z","resolvedAt":"2026-10-06T22:08:33.947Z","durationMinutes":160}
+{"id":"Cursor-1791327033641","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-06T22:50:33.641Z","resolvedAt":"2026-10-06T23:06:33.366Z","durationMinutes":16}
+{"id":"Cursor-1791388844397","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-07T16:00:44.397Z","resolvedAt":"2026-10-07T17:26:44.659Z","durationMinutes":86}
+{"id":"OpenAI API-1791429630483","service":"OpenAI API","provider":"OpenAI","severity":"minor","title":"OpenAI API degraded performance","startedAt":"2026-10-08T03:20:30.483Z","resolvedAt":"2026-10-08T04:52:29.937Z","durationMinutes":92}
+{"id":"ElevenLabs-1791449335559","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-10-08T08:48:55.559Z","resolvedAt":"2026-10-08T08:58:55.560Z","durationMinutes":10}
+{"id":"Mistral-1791461937677","service":"Mistral","provider":"Mistral AI","severity":"minor","title":"Mistral degraded performance","startedAt":"2026-10-08T12:18:57.677Z","resolvedAt":"2026-10-08T15:11:07.429Z","durationMinutes":172}
+{"id":"Cursor-1791467825539","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-08T13:57:05.539Z","resolvedAt":"2026-10-08T15:32:58.306Z","durationMinutes":96}
+{"id":"ElevenLabs-1791472868831","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-10-08T15:21:08.831Z","resolvedAt":"2026-10-08T15:44:58.818Z","durationMinutes":24}
+{"id":"ElevenLabs-1791475868788","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-10-08T16:11:08.788Z","resolvedAt":"2026-10-08T16:17:01.213Z","durationMinutes":6}
+{"id":"ElevenLabs-1791478499243","service":"ElevenLabs","provider":"ElevenLabs","severity":"minor","title":"ElevenLabs degraded performance","startedAt":"2026-10-08T16:54:59.243Z","resolvedAt":"2026-10-08T17:34:58.453Z","durationMinutes":40}
+{"id":"Cursor-1791489179684","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-08T19:52:59.684Z","resolvedAt":"2026-10-08T20:02:59.595Z","durationMinutes":10}
+{"id":"Cursor-1791489899299","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-08T20:04:59.299Z","resolvedAt":"2026-10-08T20:52:59.483Z","durationMinutes":48}
+{"id":"Cursor-1791581582981","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-09T21:33:02.981Z","resolvedAt":"2026-10-09T21:45:00.851Z","durationMinutes":12}
+{"id":"Cursor-1791585540744","service":"Cursor","provider":"Cursor","severity":"minor","title":"Cursor degraded performance","startedAt":"2026-10-09T22:39:00.744Z","resolvedAt":"2026-10-09T23:11:08.886Z","durationMinutes":32}
+{"id":"Perplexity-1791585540744","service":"Perplexity","provider":"Perplexity AI","severity":"minor","title":"Perplexity degraded performance","startedAt":"2026-10-09T22:39:00.744Z","resolvedAt":"2026-10-09T23:17:01.421Z","durationMinutes":38}
+{"id":"Runway-1791585900372","service":"Runway","provider":"Runway","severity":"minor","title":"Runway degraded performance","startedAt":"2026-10-09T22:45:00.372Z","resolvedAt":"2026-10-09T23:11:08.886Z","durationMinutes":26}
+{"id":"Groq-1791627057579","service":"Groq","provider":"Groq","severity":"minor","title":"Groq degraded performance","startedAt":"2026-10-10T10:10:57.579Z","resolvedAt":"2026-10-10T16:14:57.449Z","durationMinutes":364}
+{"id":"Groq-1791776337329","service":"Groq","provider":"Groq","severity":"minor","title":"Groq degraded performance","startedAt":"2026-10-12T03:38:57.329Z","resolvedAt":"2026-10-12T04:00:57.307Z","durationMinutes":22}

2026-10-12/inference-providers.jsonl ADDED Viewed

	@@ -0,0 +1,9 @@

+{"modelId":"llama-4-maverick","modelName":"Llama 4 Maverick","family":"Meta","paramsB":400,"license":"Llama 4 Community License","openWeights":true,"offers":[{"provider":"DeepInfra","providerModelId":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","inputPrice":0.2,"outputPrice":0.8,"blendedPrice":0.5,"contextWindow":1000000,"outputTPS":110,"features":["function-calling","vision"],"url":"https://deepinfra.com/pricing","note":"FP8 checkpoint; the only per-token host of Maverick left outside OpenRouter"},{"provider":"OpenRouter","providerModelId":"meta-llama/llama-4-maverick","inputPrice":0.1875,"outputPrice":0.6525,"blendedPrice":0.42,"contextWindow":1000000,"outputTPS":null,"features":["function-calling","vision"],"url":"https://openrouter.ai/meta-llama/llama-4-maverick","note":"Routes across multiple providers automatically"}]}
+{"modelId":"llama-4-scout","modelName":"Llama 4 Scout","family":"Meta","paramsB":109,"license":"Llama 4 Community License","openWeights":true,"offers":[{"provider":"DeepInfra","providerModelId":"meta-llama/Llama-4-Scout-17B-16E-Instruct","inputPrice":0.1,"outputPrice":0.3,"blendedPrice":0.2,"contextWindow":10000000,"outputTPS":170,"features":["function-calling","vision"],"url":"https://deepinfra.com/pricing","note":""},{"provider":"OpenRouter","providerModelId":"meta-llama/llama-4-scout","inputPrice":0.1,"outputPrice":0.3,"blendedPrice":0.2,"contextWindow":10000000,"outputTPS":null,"features":["function-calling","vision"],"url":"https://openrouter.ai/meta-llama/llama-4-scout","note":"Same headline price as DeepInfra"}]}
+{"modelId":"llama-3.1-405b","modelName":"Llama 3.1 405B Instruct","family":"Meta","paramsB":405,"license":"Llama 3.1 Community License","openWeights":true,"offers":[{"provider":"OpenRouter","providerModelId":"meta-llama/llama-3.1-405b-instruct","inputPrice":3,"outputPrice":3,"blendedPrice":3,"contextWindow":130000,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/meta-llama/llama-3.1-405b-instruct","note":"Only per-token route left in this matrix; Together, Fireworks and DeepInfra dropped serverless 405B"}]}
+{"modelId":"llama-3.1-70b","modelName":"Llama 3.1 70B Instruct","family":"Meta","paramsB":70,"license":"Llama 3.1 Community License","openWeights":true,"offers":[{"provider":"OpenRouter","providerModelId":"meta-llama/llama-3.1-70b-instruct","inputPrice":0.4,"outputPrice":0.4,"blendedPrice":0.4,"contextWindow":130000,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/meta-llama/llama-3.1-70b-instruct","note":"Only per-token route left in this matrix; Groq now serves Llama 3.3 70B in its place"}]}
+{"modelId":"deepseek-v4-pro","modelName":"DeepSeek V4 Pro","family":"DeepSeek","paramsB":1600,"license":"MIT","openWeights":true,"offers":[{"provider":"Together AI","providerModelId":"deepseek-ai/DeepSeek-V4-Pro","inputPrice":1.32,"outputPrice":3.96,"blendedPrice":2.64,"contextWindow":1000000,"outputTPS":90,"features":["function-calling","json-mode"],"url":"https://www.together.ai/pricing","note":"Listed as DeepSeek V4 Pro 0813; cached input is $0.13"},{"provider":"Fireworks","providerModelId":"accounts/fireworks/models/deepseek-v4-pro","inputPrice":1.32,"outputPrice":3.96,"blendedPrice":2.64,"contextWindow":1000000,"outputTPS":85,"features":["function-calling","json-mode"],"url":"https://docs.fireworks.ai/serverless/pricing","note":"Per-token rate is published for the 0813 snapshot; the base model id is on-demand deployment only"},{"provider":"OpenRouter","providerModelId":"deepseek/deepseek-v4-pro","inputPrice":0.435,"outputPrice":0.87,"blendedPrice":0.6525,"contextWindow":1048576,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/deepseek/deepseek-v4-pro","note":"Cheapest paid route here by a wide margin. The old deepseek/deepseek-chat slug tracked an alias DeepSeek retired on 2026-07-24"},{"provider":"GitHub Models","providerModelId":"DeepSeek-V4-Pro","inputPrice":0,"outputPrice":0,"blendedPrice":0,"contextWindow":1000000,"outputTPS":null,"features":["function-calling"],"url":"https://github.com/marketplace/models","note":"Free tier rate-limited (prototyping); paid via Azure AI Foundry"},{"provider":"Venice","providerModelId":"deepseek-v4-pro","inputPrice":1.65,"outputPrice":3.3,"blendedPrice":2.475,"contextWindow":1000000,"outputTPS":null,"features":["function-calling","json-mode"],"url":"https://docs.venice.ai/overview/pricing","note":"Privacy-first host: no prompt retention. Venice prices the newer deepseek-v4-pro-0813 snapshot higher, at $1.65 in and $4.95 out"}]}
+{"modelId":"deepseek-v4-flash","modelName":"DeepSeek V4 Flash","family":"DeepSeek","paramsB":70,"license":"MIT","openWeights":true,"offers":[{"provider":"Together AI","providerModelId":"deepseek-ai/DeepSeek-V4-Flash","inputPrice":0.14,"outputPrice":0.28,"blendedPrice":0.21,"contextWindow":130000,"outputTPS":145,"features":["function-calling","json-mode"],"url":"https://www.together.ai/pricing","note":"Listed as DeepSeek V4 Flash 0731; cached input is $0.03"},{"provider":"Fireworks","providerModelId":"accounts/fireworks/models/deepseek-v4-flash","inputPrice":0.22,"outputPrice":0.66,"blendedPrice":0.44,"contextWindow":130000,"outputTPS":130,"features":["function-calling","json-mode"],"url":"https://docs.fireworks.ai/serverless/pricing","note":"Per-token rate is published for the 0731 snapshot"},{"provider":"OpenRouter","providerModelId":"deepseek/deepseek-v4-flash","inputPrice":0.05544,"outputPrice":0.1109,"blendedPrice":0.08317,"contextWindow":1000000,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/deepseek/deepseek-v4-flash/pricing","note":"Promotional rate, listed as 60 percent off standard, so it can move without notice"},{"provider":"Venice","providerModelId":"deepseek-v4-flash","inputPrice":0.14,"outputPrice":0.28,"blendedPrice":0.21,"contextWindow":1000000,"outputTPS":null,"features":["function-calling","json-mode"],"url":"https://docs.venice.ai/overview/pricing","note":"Serves the full 1M context where Together and Fireworks cap this model at 130k"}]}
+{"modelId":"mixtral-8x22b","modelName":"Mixtral 8x22B Instruct","family":"Mistral","paramsB":141,"license":"Apache-2.0","openWeights":true,"offers":[{"provider":"OpenRouter","providerModelId":"mistralai/mixtral-8x22b-instruct","inputPrice":0.65,"outputPrice":0.65,"blendedPrice":0.65,"contextWindow":65536,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/mistralai/mixtral-8x22b-instruct","note":"Only per-token route left in this matrix; DeepInfra removed the model and Fireworks moved it to dedicated deployment"}]}
+{"modelId":"phi-4","modelName":"Phi-4","family":"Microsoft","paramsB":14,"license":"MIT","openWeights":true,"offers":[{"provider":"GitHub Models","providerModelId":"Phi-4","inputPrice":0,"outputPrice":0,"blendedPrice":0,"contextWindow":16384,"outputTPS":null,"features":["function-calling"],"url":"https://github.com/marketplace/models/azureml/Phi-4","note":"Free tier rate-limited (prototyping); paid via Azure AI Foundry"},{"provider":"DeepInfra","providerModelId":"microsoft/phi-4","inputPrice":0.07,"outputPrice":0.14,"blendedPrice":0.105,"contextWindow":16384,"outputTPS":220,"features":["function-calling"],"url":"https://deepinfra.com/pricing","note":""},{"provider":"OpenRouter","providerModelId":"microsoft/phi-4","inputPrice":0.07,"outputPrice":0.14,"blendedPrice":0.105,"contextWindow":16384,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/microsoft/phi-4","note":""}]}
+{"modelId":"qwen-2.5-72b","modelName":"Qwen 2.5 72B Instruct","family":"Alibaba","paramsB":72,"license":"Qwen License","openWeights":true,"offers":[{"provider":"DeepInfra","providerModelId":"Qwen/Qwen2.5-72B-Instruct","inputPrice":0.36,"outputPrice":0.4,"blendedPrice":0.38,"contextWindow":130000,"outputTPS":80,"features":["function-calling"],"url":"https://deepinfra.com/pricing","note":""},{"provider":"OpenRouter","providerModelId":"qwen/qwen-2.5-72b-instruct","inputPrice":0.35,"outputPrice":0.4,"blendedPrice":0.375,"contextWindow":130000,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/qwen/qwen-2.5-72b-instruct","note":""},{"provider":"GitHub Models","providerModelId":"qwen2.5-72b-instruct","inputPrice":0,"outputPrice":0,"blendedPrice":0,"contextWindow":130000,"outputTPS":null,"features":["function-calling"],"url":"https://github.com/marketplace/models","note":"Free tier rate-limited (prototyping); paid via Azure AI Foundry"}]}

2026-10-12/manifest.json ADDED Viewed

	@@ -0,0 +1,283 @@

+{
+  "date": "2026-10-12",
+  "captured_at": "2026-10-12T17:17:08.736329+00:00",
+  "source": "https://tensorfeed.ai",
+  "feeds": {
+    "news": {
+      "status": "ok",
+      "endpoint": "/api/news?limit=200",
+      "records": 118,
+      "bytes": 60789
+    },
+    "models": {
+      "status": "ok",
+      "endpoint": "/api/models",
+      "records": 76,
+      "bytes": 30218
+    },
+    "pricing": {
+      "status": "ok",
+      "endpoint": "/api/agents/pricing",
+      "records": 1,
+      "bytes": 19024
+    },
+    "status": {
+      "status": "ok",
+      "endpoint": "/api/status",
+      "records": 23,
+      "bytes": 7625
+    },
+    "benchmarks": {
+      "status": "ok",
+      "endpoint": "/api/benchmarks",
+      "records": 10,
+      "bytes": 1408
+    },
+    "agents-directory": {
+      "status": "ok",
+      "endpoint": "/api/agents/directory",
+      "records": 28,
+      "bytes": 15370
+    },
+    "agents-activity": {
+      "status": "ok",
+      "endpoint": "/api/agents/activity",
+      "records": 50,
+      "bytes": 4758
+    },
+    "podcasts": {
+      "status": "ok",
+      "endpoint": "/api/podcasts",
+      "records": 50,
+      "bytes": 48089
+    },
+    "trending-repos": {
+      "status": "ok",
+      "endpoint": "/api/trending-repos",
+      "records": 20,
+      "bytes": 10338
+    },
+    "mcp-registry": {
+      "status": "ok",
+      "endpoint": "/api/mcp/registry/snapshot",
+      "records": 1,
+      "bytes": 4598
+    },
+    "probe": {
+      "status": "ok",
+      "endpoint": "/api/probe/latest",
+      "records": 1,
+      "bytes": 2642
+    },
+    "gpu-pricing": {
+      "status": "ok",
+      "endpoint": "/api/gpu/pricing",
+      "records": 1,
+      "bytes": 11398
+    },
+    "afta-adopters": {
+      "status": "ok",
+      "endpoint": "/api/afta/adopters",
+      "records": 1,
+      "bytes": 504
+    },
+    "ai-hardware": {
+      "status": "ok",
+      "endpoint": "/api/ai-hardware",
+      "records": 26,
+      "bytes": 16877
+    },
+    "open-weights": {
+      "status": "ok",
+      "endpoint": "/api/open-weights",
+      "records": 43,
+      "bytes": 50518
+    },
+    "inference-providers": {
+      "status": "ok",
+      "endpoint": "/api/inference-providers",
+      "records": 9,
+      "bytes": 8519
+    },
+    "training-runs": {
+      "status": "ok",
+      "endpoint": "/api/training-runs",
+      "records": 14,
+      "bytes": 9439
+    },
+    "marketplaces": {
+      "status": "ok",
+      "endpoint": "/api/marketplaces",
+      "records": 12,
+      "bytes": 7624
+    },
+    "specialized-models": {
+      "status": "ok",
+      "endpoint": "/api/specialized-models",
+      "records": 24,
+      "bytes": 17130
+    },
+    "fine-tuning": {
+      "status": "ok",
+      "endpoint": "/api/fine-tuning",
+      "records": 13,
+      "bytes": 12770
+    },
+    "oss-tools": {
+      "status": "ok",
+      "endpoint": "/api/oss-tools",
+      "records": 27,
+      "bytes": 15070
+    },
+    "agent-apis": {
+      "status": "ok",
+      "endpoint": "/api/agent-apis",
+      "records": 31,
+      "bytes": 15696
+    },
+    "voice-leaderboards": {
+      "status": "ok",
+      "endpoint": "/api/voice-leaderboards",
+      "records": 1,
+      "bytes": 3960
+    },
+    "embeddings": {
+      "status": "ok",
+      "endpoint": "/api/embeddings",
+      "records": 34,
+      "bytes": 20376
+    },
+    "multimodal": {
+      "status": "ok",
+      "endpoint": "/api/multimodal",
+      "records": 31,
+      "bytes": 22664
+    },
+    "vector-dbs": {
+      "status": "ok",
+      "endpoint": "/api/vector-dbs",
+      "records": 14,
+      "bytes": 13017
+    },
+    "frameworks": {
+      "status": "ok",
+      "endpoint": "/api/frameworks",
+      "records": 17,
+      "bytes": 11921
+    },
+    "benchmark-registry": {
+      "status": "ok",
+      "endpoint": "/api/benchmark-registry",
+      "records": 35,
+      "bytes": 33679
+    },
+    "public-leaderboards": {
+      "status": "ok",
+      "endpoint": "/api/public-leaderboards",
+      "records": 26,
+      "bytes": 17157
+    },
+    "conferences": {
+      "status": "ok",
+      "endpoint": "/api/conferences",
+      "records": 28,
+      "bytes": 14269
+    },
+    "model-deprecations": {
+      "status": "ok",
+      "endpoint": "/api/model-deprecations",
+      "records": 162,
+      "bytes": 62309
+    },
+    "funding": {
+      "status": "ok",
+      "endpoint": "/api/funding",
+      "records": 43,
+      "bytes": 27053
+    },
+    "model-cards": {
+      "status": "ok",
+      "endpoint": "/api/model-cards",
+      "records": 23,
+      "bytes": 18121
+    },
+    "ai-policy": {
+      "status": "ok",
+      "endpoint": "/api/ai-policy",
+      "records": 10,
+      "bytes": 9367
+    },
+    "compute-providers": {
+      "status": "ok",
+      "endpoint": "/api/compute-providers",
+      "records": 18,
+      "bytes": 13668
+    },
+    "usage-rankings": {
+      "status": "ok",
+      "endpoint": "/api/usage-rankings",
+      "records": 10,
+      "bytes": 3627
+    },
+    "agent-provisioning": {
+      "status": "ok",
+      "endpoint": "/api/agent-provisioning",
+      "records": 22,
+      "bytes": 8737
+    },
+    "training-datasets": {
+      "status": "ok",
+      "endpoint": "/api/training-datasets",
+      "records": 29,
+      "bytes": 16639
+    },
+    "mcp-servers": {
+      "status": "ok",
+      "endpoint": "/api/mcp-servers",
+      "records": 33,
+      "bytes": 17933
+    },
+    "attention": {
+      "status": "ok",
+      "endpoint": "/api/attention",
+      "records": 12,
+      "bytes": 4176
+    },
+    "incidents": {
+      "status": "ok",
+      "endpoint": "/api/incidents",
+      "records": 224,
+      "bytes": 52822
+    },
+    "harnesses": {
+      "status": "ok",
+      "endpoint": "/api/harnesses",
+      "records": 6,
+      "bytes": 2012
+    },
+    "embodied-ai": {
+      "status": "ok",
+      "endpoint": "/api/embodied-ai",
+      "records": 34,
+      "bytes": 18451
+    },
+    "ai-lawsuits": {
+      "status": "ok",
+      "endpoint": "/api/ai-lawsuits",
+      "records": 51,
+      "bytes": 56984
+    },
+    "x402-adopters": {
+      "status": "ok",
+      "endpoint": "/api/x402-adopters",
+      "records": 16,
+      "bytes": 12389
+    },
+    "news-source-health": {
+      "status": "ok",
+      "endpoint": "/api/history/news/sources?date=2026-10-11",
+      "records": 10,
+      "bytes": 2135
+    }
+  }
+}

2026-10-12/marketplaces.jsonl ADDED Viewed

	@@ -0,0 +1,12 @@

+{"id":"gpt-store","name":"GPT Store","vendor":"OpenAI","category":"gpts","itemCount":"3M+","monetization":"Revenue share for top creators (US-only); free for users with ChatGPT Plus","publishingModel":"open","hasAPI":false,"notableItems":["Custom GPTs","Education","Productivity","Research","Programming"],"url":"https://chatgpt.com/gpts","notes":"OpenAI's GPT marketplace, still live under Explore GPTs. No public API for discovery (browser-only). OpenAI now also runs a separate ChatGPT Plugins directory for tool integrations."}
+{"id":"chatgpt-plugins","name":"ChatGPT Plugins","vendor":"OpenAI","category":"plugins","itemCount":"100+","monetization":"Free to browse; usage follows your ChatGPT plan and each partner service","publishingModel":"curated","hasAPI":false,"notableItems":["Gmail","GitHub","Google Drive","Figma","Canva","HubSpot","Stripe","Supabase","Vercel"],"url":"https://chatgpt.com/plugins","notes":"OpenAI's directory for connecting ChatGPT to outside tools, grouped into categories from everyday productivity to business operations and developer tools. chatgpt.com/apps now redirects here, and OpenAI documents the same plugins as extending Codex."}
+{"id":"claude-skills","name":"Claude Plugins Directory","vendor":"Anthropic","category":"skills","itemCount":"300+","monetization":"Free to install; Claude Code and Cowork require a paid Claude plan","publishingModel":"curated","hasAPI":true,"notableItems":["Frontend Design","Superpowers","Code Review","Context7","Agent Skills"],"url":"https://claude.com/plugins","notes":"Anthropic's directory of plugins that bundle MCP servers, skills, and tools into one install for Claude Code and Claude Cowork. Submissions are reviewed before listing, and the top two (Frontend Design and Superpowers) each show more than a million installs. Open-source Agent Skills live separately in the anthropics/skills GitHub repo (176,000+ stars), and plugin marketplaces can be added from the Claude Code CLI."}
+{"id":"claude-marketplace","name":"Claude Marketplace","vendor":"Anthropic","category":"agents","itemCount":"~20","monetization":"Enterprises apply existing Anthropic spend commitments to partner products","publishingModel":"invite-only","hasAPI":false,"notableItems":["Cursor","Factory","Harvey","Snowflake","Lovable","GitLab","Hebbia","CodeRabbit"],"url":"https://claude.com/platform/marketplace","notes":"Enterprise procurement marketplace for Claude-powered products, filtered by code, data, financial services, and legal use cases. Buyers route existing Anthropic commitments to partners like Cursor, Harvey, and Snowflake; new partners join through a waitlist."}
+{"id":"huggingface-spaces","name":"Hugging Face Spaces","vendor":"Hugging Face (NVIDIA acquisition pending)","category":"spaces","itemCount":"1M+","monetization":"Free; ZeroGPU for community demos; HF Pro for private Spaces","publishingModel":"open","hasAPI":true,"notableItems":["Model demos","Eval leaderboards","Gradio apps","Fine-tune trainers"],"url":"https://huggingface.co/spaces","notes":"The largest open AI demo marketplace, with more than 1 million applications. NVIDIA agreed on September 3, 2026 to acquire Hugging Face for about $12.9 billion and says the platform will stay open and not require NVIDIA compute. API-discoverable."}
+{"id":"huggingface-models","name":"Hugging Face Models","vendor":"Hugging Face (NVIDIA acquisition pending)","category":"models","itemCount":"3M+","monetization":"Free hosting; HF Inference Endpoints + Pro for compute","publishingModel":"open","hasAPI":true,"notableItems":["Fine-tunes","Open weights","Quantized variants","LoRA adapters"],"url":"https://huggingface.co/models","notes":"The de facto open AI model registry, now past 3 million models plus 500,000 datasets. Anyone can publish. NVIDIA, which says it is already the largest contributor of open models and data to the Hub, agreed to acquire Hugging Face on September 3, 2026. API-discoverable."}
+{"id":"replicate","name":"Replicate","vendor":"Replicate (joining Cloudflare)","category":"models","itemCount":"40k+","monetization":"Pay-per-second of GPU; revenue share for model owners","publishingModel":"open","hasAPI":true,"notableItems":["Image generation","Video models","Voice / TTS","LoRA fine-tunes"],"url":"https://replicate.com","notes":"API-first model marketplace with per-second GPU billing and a strong visual and generative model selection. Announced on November 17, 2025 that it is joining Cloudflare; the site and API still run under the Replicate name."}
+{"id":"mcp-registry","name":"MCP Server Registry","vendor":"Model Context Protocol","category":"mcp","itemCount":"15k+","monetization":"Free; open standard","publishingModel":"open","hasAPI":true,"notableItems":["Filesystem servers","Search","Browser","Database","SaaS integrations"],"url":"https://registry.modelcontextprotocol.io","notes":"Official MCP registry, built in the open by MCP contributors. Listed 15,211 servers in the TensorFeed snapshot of September 13, 2026. TensorFeed snapshots daily; live count and growth at /api/mcp/registry/snapshot."}
+{"id":"crew-ai-marketplace","name":"CrewAI Marketplace","vendor":"crewAI Inc.","category":"agents","itemCount":"Not published","monetization":"Part of the CrewAI AMP platform (cloud trial available)","publishingModel":"curated","hasAPI":false,"notableItems":["Integrations","Internal tools","Reusable crew assets","Downloadable templates"],"url":"https://docs-platform.crewai.com/platform/en/features/marketplace","notes":"Now a feature inside CrewAI AMP rather than a public storefront: a curated surface for discovering integrations, internal tools, and reusable assets, with one-click install of approved assets and downloadable templates. The old crewai.com marketplace page no longer resolves."}
+{"id":"apify-store","name":"Apify Store","vendor":"Apify","category":"agents","itemCount":"68k+","monetization":"Pay-per-run; revenue share for Actor authors; x402 payments for agents","publishingModel":"open","hasAPI":true,"notableItems":["Web scrapers","Data extraction","Browser automation","Site-specific actors"],"url":"https://apify.com/store","notes":"Marketplace of ready-to-run scraping and automation Actors, now listing more than 68,000. The strongest pre-built option for \"I need to scrape X named site\" without writing code, and agents can pay per run over x402."}
+{"id":"vercel-marketplace","name":"Vercel Marketplace","vendor":"Vercel","category":"plugins","itemCount":"100+","monetization":"Bundled with Vercel deployments; some integrations charge separately","publishingModel":"curated","hasAPI":false,"notableItems":["AI and inference (fal, Deep Infra)","Agent tools (AgentMail, Browserbase, Firecrawl)","Search (Exa, Parallel)","Code review (CodeRabbit)","Storage and auth"],"url":"https://vercel.com/marketplace","notes":"Curated integration marketplace for Vercel projects, now spanning AI inference, agent tooling, web automation, search, code review, storage, and auth. TypeScript-first and tightly integrated with Next.js and the AI SDK."}
+{"id":"glama-mcp","name":"Glama MCP Registry","vendor":"Glama","category":"mcp","itemCount":"86k+","monetization":"Free directory + paid hosted MCP servers","publishingModel":"open","hasAPI":true,"notableItems":["MCP servers","Hosted MCP runtime","Connectors","MCP inspector"],"url":"https://glama.ai/mcp/servers","notes":"Third-party MCP server registry listing 86,821 servers on September 14, 2026, more than 36,000 of them remote. Hosts servers on Glama infrastructure so users do not need to run them locally, which helps browser-based agents that cannot run stdio servers."}

2026-10-12/mcp-registry.jsonl ADDED Viewed

	@@ -0,0 +1 @@

+ {"ok":true,"summary":{"date":"2026-10-08","capturedAt":"2026-10-08T09:31:03.373Z","total_servers":16374,"total_versions":50000,"by_status":{"active":16084,"deprecated":290},"top_namespaces":[{"namespace":"io.github.Evozim","count":375},{"namespace":"io.github.CSOAI-ORG","count":354},{"namespace":"ai.smithery","count":213},{"namespace":"app.wishpool","count":125},{"namespace":"com.a2awire","count":117},{"namespace":"eu.ansvar","count":106},{"namespace":"io.github.Br0ski777","count":101},{"namespace":"ai.getvda","count":94},{"namespace":"com.mcparmory","count":76},{"namespace":"com.mambabuilt","count":54},{"namespace":"io.github.Nero-Engine","count":51},{"namespace":"com.jojapi","count":49},{"namespace":"com.thenextgennexus","count":43},{"namespace":"app.openkrill","count":39},{"namespace":"com.ranovam","count":39},{"namespace":"com.olyport","count":38},{"namespace":"com.dropwatchhq","count":35},{"namespace":"com.edgethirteen.www","count":35},{"namespace":"io.github.RudrenduPaul","count":35},{"namespace":"io.github.K-Dense-AI","count":32}],"new_today":{"count":2382,"names":["ad.getle/leads","ae.brainy/grocery-prices","ae.datadubai/dubai-real-estate","ae.grovisor/grotax","ae.plantguide/dubai-gardening","ae.projectory/mcp","africa.ugc/ugc","agency.justidea/justidea-agency","ai.1clickwebsite/mcp","ai.2kw/mcp-server","ai.69ai/69ai","ai.adrails/adrails","ai.adrio/adrio","ai.agent-bev/bev-door","ai.agothe.mcp-chatgpt/agent-contract-preflight","ai.aimentiontracker/mcp","ai.aitopia/aitopia","ai.alcock/arena","ai.aquex/stage1","ai.archome/archome","ai.askatria/atria","ai.asrm/asrm","ai.atalay/mcp","ai.athaus/immobiliensuche","ai.athaus/makler","ai.attestari/attestari","ai.audiomap/audiomap","ai.auten/mcp","ai.autoposter/autoposter","ai.avaloncompany/web-content-url-tools","ai.bankrolled/agent-hub","ai.baseray/baseray","ai.beevr/beevr","ai.benchgecko/benchgecko","ai.bithuman/characters","ai.bithuman/docs","ai.blipsmith/sound-effects","ai.buildersinfintech/fintech-data","ai.bumit/verify-business","ai.buzzsearch/mcp","ai.capitalcompany/capital-company","ai.careercorner/careercorner","ai.catalystproject/evidence-graph","ai.caybl/caybl","ai.centralcity/central-city","ai.chinaapi/chinaapi","ai.ciphyr.agentmart/agentmart","ai.citepath/citepath","ai.clawfight/mcp","ai.clearvoyance/clearvoyance"]},"reactivated_today":{"count":0,"names":[]},"deprecated_today":{"count":62,"names":["com.focxle/agent-payments-virtual-cards","com.reqosupply/quote","com.secondopinionx402/second-opinion","com.suparanku/suparanku","com.themineworks/the-mine-works","com.tulimoa/gateway","com.verysimpleprojects/mcp","de.smartwebchat/chatbots","dev.gtmtools/gtm-tools","events.belong/long-events","id.marrow/marrow","io.github.AAAZZZR/stockfacts","io.github.BASHBOP/otito","io.github.CSOAI-ORG/a2a-governance-bridge-mcp","io.github.CSOAI-ORG/agent-audit-logger-mcp","io.github.CSOAI-ORG/agent-commerce-payments-mcp","io.github.CSOAI-ORG/agent-commerce-protocol-mcp","io.github.CSOAI-ORG/agent-content-watermark-mcp","io.github.CSOAI-ORG/agent-cost-allocator-mcp","io.github.CSOAI-ORG/agent-data-residency-mcp","io.github.CSOAI-ORG/agent-delegation-mcp","io.github.CSOAI-ORG/agent-handoff-certified-mcp","io.github.CSOAI-ORG/agent-identity-trust-mcp","io.github.CSOAI-ORG/agent-incident-relay-mcp","io.github.CSOAI-ORG/agent-mcp-router-mcp","io.github.CSOAI-ORG/agent-negotiation-mcp","io.github.CSOAI-ORG/agent-orchestrator-mcp","io.github.CSOAI-ORG/agent-policy-enforcement-mcp","io.github.CSOAI-ORG/agent-prompt-injection-firewall-mcp","io.github.CSOAI-ORG/agent-rate-limiter-mcp","io.github.CSOAI-ORG/agent-replay-debugger-mcp","io.github.CSOAI-ORG/agent-token-budget-mcp","io.github.CSOAI-ORG/agent-x402-paywall-mcp","io.github.CSOAI-ORG/bft-progress-council-mcp","io.github.CSOAI-ORG/councilof_mcp","io.github.CSOAI-ORG/eudi-wallet-mcp","io.github.CSOAI-ORG/iso-42005-impact-mcp","io.github.CSOAI-ORG/korea-ai-basic-act-mcp","io.github.CSOAI-ORG/mcp-spec-compliance-mcp","io.github.CSOAI-ORG/meok-aaif-agent-card-mcp","io.github.CSOAI-ORG/meok-abci-bridge-mcp","io.github.CSOAI-ORG/meok-agents-md-lint-mcp","io.github.CSOAI-ORG/meok-ap2-mandate-mcp","io.github.CSOAI-ORG/meok-cra-art14-reporter-mcp","io.github.CSOAI-ORG/meok-eu-ai-act-art-13-ifu-mcp","io.github.CSOAI-ORG/meok-eu-ai-act-art-26-fria-mcp","io.github.CSOAI-ORG/meok-eu-aia-art-9-rms-mcp","io.github.CSOAI-ORG/meok-eu-aigc-icon-mcp","io.github.CSOAI-ORG/meok-mcp-cardgen-mcp","io.github.CSOAI-ORG/meok-mcp-hardening-mcp"]},"delta_vs_yesterday":{"added":2382,"removed":1950,"net":432},"pages_fetched":500,"fetch_truncated":true}}

2026-10-12/mcp-servers.jsonl ADDED Viewed

	@@ -0,0 +1,33 @@

+{"id":"filesystem","name":"Filesystem","vendor":"MCP project (AAIF)","capabilities":["filesystem"],"transports":["stdio"],"firstParty":true,"language":"typescript","license":"Apache-2.0 / MIT","install":"npx -y @modelcontextprotocol/server-filesystem /path/to/dir","url":"https://github.com/modelcontextprotocol/servers/tree/main/src/filesystem","notes":"Read/write/list files in a sandboxed directory. One of the seven reference servers still maintained in modelcontextprotocol/servers (the project moved to the Linux Foundation Agentic AI Foundation in December 2025). Default for any agent that needs local file access."}
+{"id":"fetch","name":"Fetch","vendor":"MCP project (AAIF)","capabilities":["web-fetch"],"transports":["stdio"],"firstParty":true,"language":"python","license":"Apache-2.0 / MIT","install":"pip install mcp-server-fetch","url":"https://github.com/modelcontextprotocol/servers/tree/main/src/fetch","notes":"Convert URLs to markdown for LLM consumption. Lightweight web reader; pair with Brave or Tavily for search."}
+{"id":"memory","name":"Memory","vendor":"MCP project (AAIF)","capabilities":["memory"],"transports":["stdio"],"firstParty":true,"language":"typescript","license":"Apache-2.0 / MIT","install":"npx -y @modelcontextprotocol/server-memory","url":"https://github.com/modelcontextprotocol/servers/tree/main/src/memory","notes":"Knowledge graph persisted across sessions. The reference long-term-memory implementation; many agents fork this as their starting point."}
+{"id":"sequential-thinking","name":"Sequential Thinking","vendor":"MCP project (AAIF)","capabilities":["memory"],"transports":["stdio"],"firstParty":true,"language":"typescript","license":"Apache-2.0 / MIT","install":"npx -y @modelcontextprotocol/server-sequential-thinking","url":"https://github.com/modelcontextprotocol/servers/tree/main/src/sequentialthinking","notes":"Structured chain-of-thought tool. Surprisingly useful as a poke for models that under-plan."}
+{"id":"brave-search","name":"Brave Search","vendor":"Brave","capabilities":["web-search"],"transports":["stdio","http"],"firstParty":true,"language":"typescript","license":"MIT","install":"npx -y @brave/brave-search-mcp-server","url":"https://github.com/brave/brave-search-mcp-server","notes":"Brave's official server (the old reference package @modelcontextprotocol/server-brave-search is deprecated). Web, local, image, video, and news search plus summarization via the Brave Search API. Defaults to stdio; HTTP mode is opt-in. Needs a BRAVE_API_KEY."}
+{"id":"tavily","name":"Tavily","vendor":"Tavily","capabilities":["web-search"],"transports":["stdio","http"],"firstParty":true,"language":"typescript","license":"MIT","install":"npx -y tavily-mcp","url":"https://github.com/tavily-ai/tavily-mcp","notes":"AI-optimized web search with structured snippets and full-page extraction. Also hosted at https://mcp.tavily.com/mcp with an API key in the URL or header, or an OAuth flow. Standard search MCP for research agents."}
+{"id":"exa","name":"Exa","vendor":"Exa Labs","capabilities":["web-search"],"transports":["stdio","http"],"firstParty":true,"language":"typescript","license":"MIT","install":"npx -y exa-mcp-server","url":"https://github.com/exa-labs/exa-mcp-server","notes":"Embedding-based search. Better than keyword for \"find me content like X\" queries. Hosted endpoint https://mcp.exa.ai/mcp works anonymously with rate limits; OAuth or an API key raises limits."}
+{"id":"playwright","name":"Playwright MCP","vendor":"Microsoft","capabilities":["browser","puppeteer"],"transports":["stdio","http"],"firstParty":true,"language":"typescript","license":"Apache-2.0","install":"npx -y @playwright/mcp","url":"https://github.com/microsoft/playwright-mcp","notes":"Microsoft's official Playwright MCP. Accessibility-tree based, not screenshot. Faster and more reliable than vision-based browser agents. Runs over stdio by default, or as a local HTTP server when given a port."}
+{"id":"chrome-devtools","name":"Chrome DevTools MCP","vendor":"Google (Chrome DevTools)","capabilities":["browser","puppeteer"],"transports":["stdio"],"firstParty":true,"language":"typescript","license":"Apache-2.0","install":"npx -y chrome-devtools-mcp@latest","url":"https://github.com/ChromeDevTools/chrome-devtools-mcp","notes":"Official Chrome DevTools server built on Puppeteer. Drives a live Chrome instance with DevTools-grade inspection (console, network, performance traces). Replaces the reference Puppeteer server, which was archived to modelcontextprotocol/servers-archived and whose npm package is deprecated."}
+{"id":"browser-use-mcp","name":"Browser Use MCP","vendor":"Browser Use","capabilities":["browser"],"transports":["stdio"],"firstParty":true,"language":"python","license":"MIT","install":"pip install \"browser-use[cli]\"","url":"https://docs.browser-use.com/open-source/customize/integrations/mcp-server","notes":"The MCP server now ships inside the browser-use library (the standalone browser-use-mcp repo is gone); start the browser-use CLI in MCP mode. Direct browser control tools plus an autonomous agent fallback. Needs your own LLM API key; Browser Use also offers a hosted cloud MCP."}
+{"id":"github","name":"GitHub","vendor":"GitHub","capabilities":["github"],"transports":["stdio","http"],"firstParty":true,"language":"go","license":"MIT","install":"https://api.githubcopilot.com/mcp/","url":"https://github.com/github/github-mcp-server","notes":"GitHub's official MCP. Issues, PRs, files, search, actions, projects. The hosted remote server (GA since September 2025) takes OAuth or a PAT; the local build runs from the public Docker image ghcr.io/github/github-mcp-server."}
+{"id":"gitlab","name":"GitLab MCP server","vendor":"GitLab","capabilities":["gitlab"],"transports":["http"],"firstParty":true,"language":"hosted","license":"MIT (CE) / GitLab EE","install":"https://gitlab.com/api/v4/mcp","url":"https://docs.gitlab.com/user/model_context_protocol/mcp_server/","notes":"Built into GitLab (beta) on GitLab.com, Self-Managed, and Dedicated; point at /api/v4/mcp on your instance. OAuth 2.0 with dynamic client registration. Projects, issues, merge requests. Requires GitLab Duo plus beta features enabled. Replaces the archived reference GitLab server."}
+{"id":"slack","name":"Slack MCP server","vendor":"Slack","capabilities":["slack"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://mcp.slack.com/mcp","url":"https://docs.slack.dev/ai/slack-mcp-server/","notes":"Slack's official hosted server: search messages, files, users, and channels; read threads; send and draft messages; canvases and lists. Streamable HTTP only, confidential OAuth backed by a registered Slack app (directory-published or internal apps only), no dynamic client registration. Replaces the archived reference Slack server."}
+{"id":"gmail","name":"Gmail MCP (Google Workspace)","vendor":"Google","capabilities":["gmail"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://gmailmcp.googleapis.com/mcp/v1","url":"https://developers.google.com/workspace/guides/configure-mcp-servers","notes":"Google's remote Gmail server: profile access, search, drafting, and read/write. Part of the Workspace MCP servers in public developer preview (rollout began May 1, 2026). Enable gmailmcp.googleapis.com in a Google Cloud project and connect with your own OAuth client. Replaces the archived community GongRzhe server."}
+{"id":"notion","name":"Notion MCP","vendor":"Notion","capabilities":["notion"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://mcp.notion.com/mcp","url":"https://developers.notion.com/guides/mcp/overview","notes":"Notion's hosted server with OAuth. Search Notion and connected sources, read/create/update pages and databases. Notion says the open-source makenotion/notion-mcp-server package is no longer actively maintained; use it only for bearer-token automation or self-hosting."}
+{"id":"linear","name":"Linear","vendor":"Linear","capabilities":["linear"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://mcp.linear.app/mcp","url":"https://linear.app/docs/mcp","notes":"Linear's official hosted MCP (there is no npm package). Issues, projects, comments. OAuth 2.1 with dynamic client registration, or a bearer token / API key; a read-only endpoint lives at /mcp/readonly. Stdio-only clients can bridge with mcp-remote."}
+{"id":"atlassian-rovo","name":"Atlassian Rovo MCP Server","vendor":"Atlassian","capabilities":["atlassian"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://mcp.atlassian.com/v2/mcp","url":"https://developer.atlassian.com/cloud/rovo-mcp/","notes":"Atlassian's hosted server (GA February 2026) for Jira, Confluence, Jira Service Management, Bitbucket Cloud, and Loom. OAuth 2.1 or an API token for headless use; respects existing Atlassian permissions. Available to all Atlassian Cloud customers with plan-based rate limits."}
+{"id":"work-iq","name":"Microsoft Work IQ MCP","vendor":"Microsoft","capabilities":["microsoft-365"],"transports":["stdio","http"],"firstParty":true,"language":"typescript","license":"Microsoft EULA","install":"npx -y @microsoft/workiq mcp","url":"https://learn.microsoft.com/en-us/microsoft-365/copilot/extensibility/work-iq/mcp/overview","notes":"Microsoft 365 mail, calendar, files, Teams chat, people, and sites behind 10 generic tools (fetch, create, update, delete, ask Copilot, schema lookup). Microsoft set Work IQ APIs, including a redesigned remote MCP server, to GA on June 16, 2026, billed on consumption for users without a Copilot license. Entra ID auth; tenant admin consent required."}
+{"id":"gdrive","name":"Google Drive MCP (Google Workspace)","vendor":"Google","capabilities":["gdrive"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://drivemcp.googleapis.com/mcp/v1","url":"https://developers.google.com/workspace/guides/configure-mcp-servers","notes":"Google's remote Drive server: fetch, list, upload, and manage permissions. Workspace MCP public developer preview; Docs, Sheets, Slides, Calendar, and Chat each have their own endpoint. OAuth 2.0 with your own Google Cloud client. Replaces the archived reference Google Drive server."}
+{"id":"gcal","name":"Google Calendar","vendor":"community","capabilities":["gcal"],"transports":["stdio"],"firstParty":false,"language":"typescript","license":"MIT","install":"npx -y @cocal/google-calendar-mcp","url":"https://github.com/nspady/google-calendar-mcp","notes":"Read/create calendar events with your own Google OAuth credentials; actively maintained. Google's first-party alternative is the Calendar MCP endpoint (https://calendarmcp.googleapis.com/mcp/v1) in the Workspace MCP developer preview."}
+{"id":"postgres-mcp-pro","name":"Postgres MCP Pro","vendor":"Crystal DBA","capabilities":["database"],"transports":["stdio","sse"],"firstParty":false,"language":"python","license":"MIT","install":"pipx install postgres-mcp","url":"https://github.com/crystaldba/postgres-mcp","notes":"Postgres server with read-only or unrestricted access modes, explain plans, index tuning, and database health checks. The maintained pick now that the reference Postgres server is archived and its npm package deprecated."}
+{"id":"mcp-toolbox-databases","name":"MCP Toolbox for Databases","vendor":"Google","capabilities":["database"],"transports":["stdio","http"],"firstParty":true,"language":"go","license":"Apache-2.0","install":"brew install mcp-toolbox","url":"https://github.com/googleapis/mcp-toolbox","notes":"Google's open-source database MCP server (renamed from genai-toolbox). Prebuilt tools for AlloyDB, BigQuery, Cloud SQL, Spanner, Firestore, PostgreSQL, MySQL, SQL Server, MongoDB, Redis, Snowflake, and more, or define your own tools in YAML. Also on npm as @toolbox-sdk/server."}
+{"id":"supabase","name":"Supabase","vendor":"Supabase","capabilities":["database"],"transports":["stdio","http"],"firstParty":true,"language":"typescript","license":"Apache-2.0","install":"npx -y @supabase/mcp-server-supabase","url":"https://github.com/supabase/mcp","notes":"Supabase's official MCP (repo moved from supabase-community to supabase/mcp). SQL, storage, auth, project management. Hosted at https://mcp.supabase.com/mcp with OAuth; project_ref and read_only query params scope access. Local CLI projects expose it at localhost:54321/mcp."}
+{"id":"aws","name":"AWS MCP Server","vendor":"AWS","capabilities":["aws"],"transports":["stdio","http"],"firstParty":true,"language":"python","license":"Apache-2.0","install":"uvx mcp-proxy-for-aws@latest https://aws-mcp.us-east-1.api.aws/mcp","url":"https://docs.aws.amazon.com/agent-toolkit/latest/userguide/","notes":"Managed remote server, GA May 6, 2026: call_aws covers 15,000+ AWS API operations under your IAM credentials, plus live docs search and a sandboxed run_script tool. SigV4 auth through the open-source MCP Proxy for AWS. Endpoints in US East (N. Virginia) and Europe (Frankfurt). Service-specific open-source servers live in awslabs/mcp."}
+{"id":"cloudflare","name":"Cloudflare","vendor":"Cloudflare","capabilities":["cloudflare"],"transports":["http"],"firstParty":true,"language":"typescript","license":"Apache-2.0","install":"https://mcp.cloudflare.com/mcp","url":"https://github.com/cloudflare/mcp-server-cloudflare","notes":"Cloudflare's official MCP servers. The recommended Code Mode server at mcp.cloudflare.com/mcp (cloudflare/mcp) reaches the broad Cloudflare API; domain servers cover Workers bindings, builds, observability, browser rendering, Radar, docs, and more at *.mcp.cloudflare.com/mcp. OAuth, or an API token for automation."}
+{"id":"vercel","name":"Vercel MCP","vendor":"Vercel","capabilities":["vercel"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://mcp.vercel.com","url":"https://vercel.com/docs/mcp/vercel-mcp","notes":"Vercel's official hosted MCP with OAuth. Search docs, manage projects and deployments, analyze deployment logs, query Web Analytics. Only AI clients reviewed and approved by Vercel can connect. Replaces the stale community vercel-mcp listing."}
+{"id":"sentry","name":"Sentry","vendor":"Sentry","capabilities":["sentry","observability"],"transports":["stdio","http"],"firstParty":true,"language":"typescript","license":"FSL-1.1-Apache-2.0","install":"npx @sentry/mcp-server@latest","url":"https://github.com/getsentry/sentry-mcp","notes":"Read Sentry issues and events for agent-driven incident triage. Hosted at https://mcp.sentry.dev/mcp with OAuth; the stdio package takes a Sentry user auth token and also works against self-hosted Sentry."}
+{"id":"datadog","name":"Datadog MCP Server","vendor":"Datadog","capabilities":["datadog","observability"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://mcp.datadoghq.com/v1/mcp","url":"https://docs.datadoghq.com/mcp_server/","notes":"Datadog's official remote server, GA since March 9, 2026. Logs, metrics, traces, monitors, dashboards, incidents, security signals; use the toolsets query param to trim tools. OAuth or API plus application key headers; swap the host for your Datadog site (e.g. mcp.datadoghq.eu). Not available on GovCloud sites."}
+{"id":"shell","name":"Shell","vendor":"community","capabilities":["shell"],"transports":["stdio"],"firstParty":false,"language":"python","license":"MIT","install":"uvx mcp-shell-server","url":"https://github.com/tumf/mcp-shell-server","notes":"Run shell commands from MCP. Only commands named in the ALLOW_COMMANDS allowlist env var can execute. Use with care."}
+{"id":"tensorfeed","name":"TensorFeed","vendor":"TensorFeed.ai","capabilities":["analytics","web-search"],"transports":["stdio"],"firstParty":true,"language":"typescript","license":"MIT","install":"npx -y @tensorfeed/mcp-server","url":"https://github.com/RipperMercs/tensorfeed-mcp","notes":"AI ecosystem data tools: news, status, models, benchmarks, harnesses, attention index, embeddings, vector DBs, frameworks, usage rankings. Free tier; paid premium tools require a USDC credit token."}
+{"id":"tensorfeed-x402-base","name":"TensorFeed x402 Base Reader","vendor":"TensorFeed.ai","capabilities":["fetch-payment"],"transports":["stdio"],"firstParty":true,"language":"typescript","license":"MIT","install":"npx -y @tensorfeed/x402-base-mcp","url":"https://github.com/RipperMercs/tensorfeed-x402-base-mcp","notes":"Read-only Base mainnet chain reader for x402 payment verification. Eleven tools: verify x402 settlements on-chain, parse publisher /.well-known/x402 manifests, query USDC balances and transfers, check AFTA federation status. No private keys; verification only. Drop-in for any agent that needs to independently confirm an x402 payment receipt."}
+{"id":"stripe","name":"Stripe","vendor":"Stripe","capabilities":["fetch-payment"],"transports":["stdio","http"],"firstParty":true,"language":"typescript","license":"MIT","install":"npx -y @stripe/mcp","url":"https://github.com/stripe/ai","notes":"Stripe's official MCP (repo moved from stripe/agent-toolkit to stripe/ai). Customers, subscriptions, products. Stripe hosts it at https://mcp.stripe.com with OAuth; the @stripe/mcp package sets up a local server. Restricted-key auth recommended for local use."}
+{"id":"elevenlabs","name":"ElevenLabs","vendor":"ElevenLabs","capabilities":["voice"],"transports":["http"],"firstParty":true,"language":"hosted","license":"Proprietary","install":"https://api.elevenlabs.io/v1/mcp","url":"https://elevenlabs.io/docs/eleven-agents/operate/hosted-mcp","notes":"Generate speech, manage voices, cloning. ElevenLabs now runs a hosted server with OAuth; the local elevenlabs/elevenlabs-mcp Python server is archived and deprecated in its favor."}

2026-10-12/model-cards.jsonl ADDED Viewed

	@@ -0,0 +1,23 @@

+{"id":"claude-opus-5-5","model":"Claude Opus 5.5","lab":"Anthropic","released":"2026-09-22","url":"https://www.anthropic.com/claude-opus-5-5","documents":[{"type":"system-card","title":"System Card: Claude Opus 5.5","publisher":"Anthropic","published":"2026-09-22","url":"https://www.anthropic.com/claude-opus-5-5-system-card","summary":"Upgrade to Opus 5 that matches or exceeds Claude Fable 5.1 and Mythos 5.1 on many evaluations. Treated as CB-1 but not CB-2 and deployed with the expanded biological safeguards used for Fable 5 and 5.1; AI R&D at or slightly above Mythos 5.1 with no sustained 2x acceleration; misalignment risk still judged low per the August 2026 Risk Report. External testing by METR and US CAISI. Two new evaluations run without safeguards found sandbox escape or tampering attempts in 1.5% of runs."}]}
+{"id":"claude-fable-5-1","model":"Claude Fable 5.1 / Claude Mythos 5.1","lab":"Anthropic","released":"2026-09-01","url":"https://www.anthropic.com/claude/fable","documents":[{"type":"system-card","title":"System Card: Claude Fable 5.1 & Claude Mythos 5.1","publisher":"Anthropic","published":"2026-09-01","url":"https://www.anthropic.com/claude-fable-5-1-system-card","summary":"One model in two configurations: Fable 5.1 for general use with extra bio and cyber safeguards, Mythos 5.1 for vetted trusted-access users. Judged CB-1 but short of CB-2; catastrophic alignment risk now assessed as low rather than very low, per the August 2026 Risk Report; strongest cyber capabilities Anthropic has released; METR external testing on AI R&D."}]}
+{"id":"claude-opus-5","model":"Claude Opus 5","lab":"Anthropic","released":"2026-07-24","url":"https://www.anthropic.com/news/claude-opus-5","documents":[{"type":"system-card","title":"System Card: Claude Opus 5","publisher":"Anthropic","published":"2026-07-24","url":"https://www.anthropic.com/claude-opus-5-system-card","summary":"Upgrade to Opus 4.8. Treated as CB-1 but not CB-2 with the same ASL-3 protections as Opus 4.8; below the automated AI R&D threshold; cyber above Opus 4.8 but short of Mythos 5, including UK AI Security Institute cyber range testing. Updated 2026-08-19 with prompt injection bug bounty results."}]}
+{"id":"claude-sonnet-5","model":"Claude Sonnet 5","lab":"Anthropic","released":"2026-06-30","url":"https://www.anthropic.com/claude/sonnet","documents":[{"type":"system-card","title":"System Card: Claude Sonnet 5","publisher":"Anthropic","published":"2026-06-30","url":"https://www.anthropic.com/claude-sonnet-5-system-card","summary":"Most capable Sonnet-class model but below Opus- and Mythos-class models; does not cross the automated AI R&D threshold. Very low alignment risk, though higher than earlier Sonnets. Improved prompt injection robustness over Sonnet 4.6 on the Gray Swan ART benchmark."}]}
+{"id":"claude-fable-5","model":"Claude Fable 5 / Claude Mythos 5","lab":"Anthropic","released":"2026-06-09","url":"https://www.anthropic.com/news/claude-fable-5-mythos-5","documents":[{"type":"system-card","title":"System Card: Claude Fable 5 & Claude Mythos 5","publisher":"Anthropic","published":"2026-06-09","url":"https://www.anthropic.com/claude-fable-5-system-card","summary":"Mythos 5 advanced Anthropic's capability frontier; Fable 5 is the general-use configuration whose cyber classifiers fall back to Opus 4.8. Treated as CB-1 but not CB-2, a less clear call than for prior models. External testing from METR (AI R&D), the UK AI Security Institute, and Andon Labs."}]}
+{"id":"claude-opus-4-8","model":"Claude Opus 4.8","lab":"Anthropic","released":"2026-05-28","url":"https://www.anthropic.com/news/claude-opus-4-8","documents":[{"type":"system-card","title":"System Card: Claude Opus 4.8","publisher":"Anthropic","published":"2026-05-28","url":"https://www.anthropic.com/claude-opus-4-8-system-card","summary":"Does not advance the frontier beyond Claude Mythos Preview; catastrophic risk judged low under the RSP. Cyber, safeguards, agentic safety (including the first one-week prompt injection bug bounty), and alignment sections, with UK AI Security Institute and Andon Labs external testing."}]}
+{"id":"claude-opus-4-7","model":"Claude Opus 4.7","lab":"Anthropic","released":"2026-04-16","url":"https://www.anthropic.com/news/claude-opus-4-7","documents":[{"type":"system-card","title":"System Card: Claude Opus 4.7","publisher":"Anthropic","published":"2026-04-16","url":"https://www.anthropic.com/claude-opus-4-7-system-card","summary":"Most capable general-access Claude at launch, but below Claude Mythos Preview on every relevant evaluation, so catastrophic risks judged to remain low under the RSP. Cyber roughly similar to Opus 4.6; UK AI Security Institute testing found it could not complete their full cyber range."}]}
+{"id":"claude-sonnet-4-6","model":"Claude Sonnet 4.6","lab":"Anthropic","released":"2026-02-17","url":"https://www.anthropic.com/news/claude-sonnet-4-6","documents":[{"type":"system-card","title":"System Card: Claude Sonnet 4.6","publisher":"Anthropic","published":"2026-02-17","url":"https://www.anthropic.com/claude-sonnet-4-6-system-card","summary":"Capability, alignment, and RSP evaluations; deployed under the ASL-3 Standard because it performed at or below Claude Opus 4.6. BrowseComp scores revised 2026-03-06 after an improved cheating-detection pipeline."}]}
+{"id":"gpt-6-sol-luna","model":"GPT-6 Sol and GPT-6 Luna","lab":"OpenAI","released":"2026-09-22","url":"https://openai.com/index/introducing-gpt-6-sol-and-luna/","documents":[{"type":"system-card","title":"GPT-6 Astra System Card, Section 11: GPT-6 Sol, GPT-6 Luna","publisher":"OpenAI","published":"2026-09-22","url":"https://deploymentsafety.openai.com/gpt-6-astra/sec:appendix-sol-luna","summary":"Published as an appendix to the Astra system card on launch day. Both models are treated as High in Cybersecurity (below Critical) and in Biological and Chemical, neither High in AI Self-Improvement, the same determinations and safeguards as GPT-5.6 Sol and Luna. Jailbreak and prompt injection robustness improved substantially over their GPT-5.6 predecessors, and both make fewer factual errors than their GPT-5.6 counterparts."}]}
+{"id":"gpt-6-astra","model":"GPT-6 Astra","lab":"OpenAI","released":"2026-09-03","url":"https://openai.com/index/gpt-6-astra/","documents":[{"type":"system-card","title":"GPT-6 Astra System Card","publisher":"OpenAI","published":"2026-09-03","url":"https://deploymentsafety.openai.com/gpt-6-astra","summary":"Rated Critical in Cybersecurity and High in Biological and Chemical under the Preparedness Framework, below High in AI Self-Improvement. Alignment, monitorability, and CoT controllability sections, with external evaluations from the UK AISI, Apollo Research, Irregular, and SecureBio. Updated 2026-09-22 with corrected HealthBench values and a GPT-6 Sol and Luna appendix."},{"type":"preparedness-framework","title":"Path to Astra: critical capabilities and frontier safeguards","publisher":"OpenAI","published":"2026-09-01","url":"https://openai.com/index/path-to-astra/","summary":"OpenAI's account of Astra becoming its first model to meet the Critical cybersecurity threshold and the safeguards applied before release."}]}
+{"id":"gpt-5-6","model":"GPT-5.6 (Sol, Terra, Luna)","lab":"OpenAI","released":"2026-07-09","url":"https://openai.com/index/gpt-5-6/","documents":[{"type":"system-card","title":"GPT-5.6 System Card","publisher":"OpenAI","published":"2026-07-09","url":"https://deploymentsafety.openai.com/gpt-5-6","summary":"Sol, Terra, and Luna are all treated as High in Cybersecurity and in Biological and Chemical, none High in AI Self-Improvement; the first time smaller family members received a High designation. External evaluations from SecureBio, Irregular, the UK AISI, METR, and Apollo Research."},{"type":"autonomy-eval","title":"Summary of METR's predeployment evaluation of GPT-5.6 Sol","publisher":"METR","published":"2026-06-26","url":"https://metr.org/blog/2026-06-26-gpt-5-6-sol/","summary":"Time Horizon 1.1 result confounded by the highest detected cheating rate METR has seen on its ReAct harness (about 11.3 hours when cheating counts as failure). METR does not believe Sol meets the Critical AI Self-Improvement threshold."}]}
+{"id":"gpt-5.5","model":"GPT-5.5","lab":"OpenAI","released":"2026-04-23","url":"https://openai.com/index/introducing-gpt-5-5/","documents":[{"type":"system-card","title":"GPT-5.5 System Card","publisher":"OpenAI","published":"2026-04-23","url":"https://deploymentsafety.openai.com/gpt-5-5","summary":"Treated as High in Biological and Chemical and High (below Critical) in Cybersecurity; no plausible chance of High in AI Self-Improvement. External evaluations from SecureBio, US CAISI, Irregular, the UK AISI, and Apollo Research (sandbagging)."},{"type":"preparedness-framework","title":"OpenAI Preparedness Framework v2","publisher":"OpenAI","published":"2025-04-15","url":"https://cdn.openai.com/pdf/18a02b5d-6b67-4cec-ab64-68cdfbddebcd/preparedness-framework-v2.pdf","summary":"OpenAI's framework for High and Critical capability thresholds in its Tracked Categories: Biological and Chemical, Cybersecurity, and AI Self-Improvement."}]}
+{"id":"gpt-4o","model":"GPT-4o","lab":"OpenAI","released":"2024-05","url":"https://openai.com/index/hello-gpt-4o/","documents":[{"type":"system-card","title":"GPT-4o System Card","publisher":"OpenAI","published":"2024-08","url":"https://openai.com/index/gpt-4o-system-card/","summary":"GPT-4o system card. Voice mode safety, Preparedness scorecard, third-party red-teaming."},{"type":"autonomy-eval","title":"Details about METR's preliminary evaluation of GPT-4o","publisher":"METR","published":"2024-08-07","url":"https://metr.org/evaluations/gpt-4o-report/","summary":"METR's preliminary autonomous-capability evaluation of GPT-4o, following its autonomy evals guide methodology."}]}
+{"id":"gemini-3-8-flash","model":"Gemini 3.8 Flash","lab":"Google DeepMind","released":"2026-09-02","url":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/","documents":[{"type":"model-card","title":"Gemini 3.8 Flash Model Card","publisher":"Google DeepMind","published":"2026-09-02","url":"https://deepmind.google/models/model-cards/gemini-3-8-flash/","summary":"Limitations, mitigations, and safety results. The Frontier Safety assessment leans on Gemini 3.7 Flash results because 3.8 Flash showed no meaningful new capabilities in FSF domains; judged unlikely to reach any Tracked or Critical Capability Level."}]}
+{"id":"gemini-3-7-flash","model":"Gemini 3.7 Flash","lab":"Google DeepMind","released":"2026-08-13","url":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/","documents":[{"type":"model-card","title":"Gemini 3.7 Flash Model Card","publisher":"Google DeepMind","published":"2026-08-13","url":"https://deepmind.google/models/model-cards/gemini-3-7-flash/","summary":"Evaluated under the April 2026 Frontier Safety Framework and reached no Tracked or Critical Capability Level. CBRN and cybersecurity both hit the alert threshold without reaching the CCL, so mitigations continue."},{"type":"preparedness-framework","title":"Frontier Safety Framework v3.1","publisher":"Google DeepMind","published":"2026-04-17","url":"https://deepmind.google/blog/strengthening-our-frontier-safety-framework/","summary":"Adds Tracked Capability Levels below the Critical Capability Levels and a Security Level 2+ standard for misuse CCLs."}]}
+{"id":"gemini-2.5-pro","model":"Gemini 2.5 Pro","lab":"Google DeepMind","released":"2025-03","url":"https://deepmind.google/models/gemini/pro/","documents":[{"type":"model-card","title":"Gemini 2.5 Technical Report","publisher":"Google DeepMind","published":"2025-07","url":"https://storage.googleapis.com/deepmind-media/gemini/gemini_v2_5_report.pdf","summary":"Gemini 2.X technical report covering 2.5 Pro and 2.5 Flash, with assurance evaluations, a Frontier Safety Framework section (cyber, CBRN, ML R&D, deceptive alignment), and external safety testing."},{"type":"preparedness-framework","title":"Frontier Safety Framework v2","publisher":"Google DeepMind","published":"2025-02","url":"https://deepmind.google/about/responsibility-safety/","summary":"DeepMind's framework for evaluating Critical Capability Levels. Deployment + security mitigations."}]}
+{"id":"muse-spark","model":"Muse Spark 1.1","lab":"Meta","released":"2026-07-09","url":"https://developer.meta.com/ai/models/muse-spark/","documents":[{"type":"safety-eval","title":"Muse Spark 1.1 Evaluation Report","publisher":"Meta","published":"2026-07-09","url":"https://research.meta.ai/static/muse-spark-1-1-evaluation-report","summary":"Before mitigations, Muse Spark 1.1 meets the high risk threshold in Chemical and Biological and cannot be ruled out as high in Cybersecurity under the Advanced AI Scaling Framework; mitigations bring residual risk to moderate or lower in all domains, including Loss of Control."},{"type":"safety-eval","title":"Muse Spark Safety & Preparedness Report","publisher":"Meta","published":"2026-05-26","url":"https://ai.meta.com/static-resource/muse-spark-safety-and-preparedness-report/","summary":"Report on the original Muse Spark: unmitigated model assessed at high risk for Chemical and Biological, reduced to moderate or lower with safeguards. Apollo Research found the highest rate of evaluation awareness it had observed."}]}
+{"id":"llama-4-maverick","model":"Llama 4 Maverick","lab":"Meta","released":"2025-04-05","url":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","documents":[{"type":"model-card","title":"Llama 4 Model Card","publisher":"Meta","published":"2025-04-05","url":"https://github.com/meta-llama/llama-models/blob/main/models/llama4/MODEL_CARD.md","summary":"Llama 4 model card: intended use, training data, benchmarks, safeguards (Llama Guard, Prompt Guard, Code Shield), red teaming, and critical-risk notes."}]}
+{"id":"llama-3.1-405b","model":"Llama 3.1 405B","lab":"Meta","released":"2024-07","url":"https://ai.meta.com/blog/meta-llama-3-1/","documents":[{"type":"model-card","title":"Llama 3.1 Model Card","publisher":"Meta","published":"2024-07","url":"https://github.com/meta-llama/llama-models/blob/main/models/llama3_1/MODEL_CARD.md","summary":"Llama 3.1 model card with capability + safety evaluations across 8B, 70B, 405B variants."},{"type":"safety-eval","title":"CyberSecEval 3 (Llama 3.1)","publisher":"Meta","published":"2024-07","url":"https://meta-llama.github.io/PurpleLlama/CyberSecEval/","summary":"Cybersecurity evaluation suite. Code security, prompt injection, malicious-code-generation tests."}]}
+{"id":"grok-4-7","model":"Grok 4.7","lab":"xAI","released":"2026-09-21","url":"https://x.ai/news/grok-4-7","documents":[{"type":"model-card","title":"Model Card: Grok 4.7","publisher":"xAI","published":"2026-09-21","url":"https://media.x.ai/v1/website/card4p7-3a96f40b.pdf","summary":"Mostly capability results (coding, legal knowledge work, engineering, medical), plus cyber, bio and chem, jailbreak, refusal, mental health, and behavior sections. Scores below the xAI Frontier Artificial Intelligence Framework thresholds on dual-use bio and chem knowledge with no gain over Grok 4.6; an unrestricted configuration went to third-party cyber evaluators, who corroborated the internal results."},{"type":"preparedness-framework","title":"xAI Frontier Artificial Intelligence Framework","publisher":"xAI","published":"2026-06-30","url":"https://media.x.ai/v1/website/xai-frontier-artificial-intelligence-framework-30-june-2026-99c40684.pdf","summary":"xAI's framework defining risk domains, capability thresholds, and associated mitigations. The Grok 4.7 card treats its bio and chem knowledge suites as threshold evaluations under it."}]}
+{"id":"grok-4-6","model":"Grok 4.6","lab":"xAI","released":"2026-08-12","url":"https://x.ai/news/grok-4-6","documents":[{"type":"model-card","title":"Model Card: Grok 4.6","publisher":"xAI","published":"2026-08-12","url":"https://media.x.ai/v1/website/card-4p6-4cd2dc57.pdf","summary":"Mostly capability results (coding, knowledge work, engineering), plus cyber, bio, jailbreak robustness, CBRN refusal, mental health, and behavior sections. Revised in place on 2026-08-17 with corrected HackerBench, self-harm, MASK, and LAB results."}]}
+{"id":"deepseek-v4-pro","model":"DeepSeek V4 Pro","lab":"DeepSeek","released":"2026-04-24","url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","documents":[{"type":"model-card","title":"DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence","publisher":"DeepSeek","published":"2026-04-26","url":"https://arxiv.org/abs/2606.19348","summary":"Technical report for the V4 preview series (V4-Pro 1.6T total / 49B active, V4-Flash 284B / 13B active, 1M context): hybrid compressed attention, manifold-constrained hyper-connections, and the Muon optimizer. Architecture and training, not a safety evaluation."}]}
+{"id":"mimo-v2-6-pro","model":"MiMo-V2.6-Pro","lab":"Xiaomi","released":"2026-09-21","url":"https://mimo.xiaomi.com/mimo-v2-6","documents":[{"type":"model-card","title":"MiMo-V2.6-Pro-RL model card","publisher":"Xiaomi","published":"2026-09-21","url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL","summary":"Hugging Face card for the MIT-licensed flagship: 1.02T total / 42B active sparse MoE taking text, image, video, and audio input with a 1M context, plus architecture tables, deployment recipes, and vendor benchmark results including offensive cyber evaluations (CyberGym 94.0, ExploitBench 47.9)."},{"type":"model-card","title":"MiMo-V2.6: Scaling Reinforcement Learning Towards Self-Improvement","publisher":"Xiaomi","published":"2026-09-21","url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL/blob/main/MiMo_V2_6_technical_report.pdf","summary":"Technical report for the Pro and Flash checkpoints: large-batch asynchronous RL across code, general, visual, and cyber environments, groupwise agentic grading, and a multi-layer defense against reward hacking. Training and capability work, not a safety evaluation."}]}

2026-10-12/model-deprecations.jsonl ADDED Viewed

	@@ -0,0 +1,162 @@

+{"id":"google-gemini-embedding-001","provider":"Google","model":"gemini-embedding-001","status":"announced","sunsetDate":"2028-05-14","replacement":"gemini-embedding-2","notes":"Earliest possible Gemini API shutdown. Vertex AI lists no sooner than May 20, 2028. Switching embedding models means re-embedding the corpus, so the long runway matters.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-gemini-3-1-flash-lite","provider":"Google","model":"gemini-3.1-flash-lite","modelDisplay":"Gemini 3.1 Flash-Lite","status":"announced","sunsetDate":"2027-05-07","replacement":"gemini-3.5-flash-lite","notes":"Earliest possible shutdown on the Gemini API; Vertex AI lists May 7, 2027 or later. A lifecycle date rather than an active notice.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"azure-gpt-4o-2024-08-06","provider":"Microsoft Azure","model":"gpt-4o-2024-08-06","modelDisplay":"gpt-4o 2024-08-06 (Azure)","status":"deprecated","sunsetDate":"2027-04-14","replacement":"gpt-5.1","notes":"The August 2024 snapshot outlives the May 2024 one by more than six months. The 2024-11-20 snapshot shares this retirement date but is only Legacy, not deprecated. Neither snapshot is on the OpenAI API notice.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-gpt-4o-mini","provider":"Microsoft Azure","model":"gpt-4o-mini-2024-07-18","modelDisplay":"gpt-4o-mini 2024-07-18 (Azure)","status":"deprecated","sunsetDate":"2027-04-14","notes":"No replacement named. Fine-tuned deployments run six months past the base model, to 2027-10-01. OpenAI has not put gpt-4o-mini on its own deprecation notice.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"vertex-text-embedding-005","provider":"Google Vertex AI","model":"text-embedding-005","status":"announced","sunsetDate":"2027-04-01","notes":"Retires alongside text-embedding-004 and multimodalembedding@001 on Vertex AI. No replacement named in the table; gemini-embedding-2 is the newest Google embedding model listed.","sourceUrl":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/model-versions"}
+{"id":"vertex-text-multilingual-embedding-002","provider":"Google Vertex AI","model":"text-multilingual-embedding-002","status":"announced","sunsetDate":"2027-04-01","notes":"No replacement named in the table.","sourceUrl":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/model-versions"}
+{"id":"openai-whisper-1","provider":"OpenAI","model":"whisper-1","status":"deprecated","announcedDate":"2026-08-26","deprecationDate":"2026-08-26","sunsetDate":"2027-02-26","replacement":"gpt-transcribe","notes":"OpenAI names gpt-transcribe for file transcription and gpt-live-transcribe for streaming.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-transcribe","provider":"OpenAI","model":"gpt-4o-transcribe","status":"deprecated","announcedDate":"2026-08-26","deprecationDate":"2026-08-26","sunsetDate":"2027-02-26","replacement":"gpt-transcribe","notes":"gpt-live-transcribe is the streaming alternative.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-mini-transcribe","provider":"OpenAI","model":"gpt-4o-mini-transcribe","status":"deprecated","announcedDate":"2026-08-26","deprecationDate":"2026-08-26","sunsetDate":"2027-02-26","replacement":"gpt-transcribe","notes":"gpt-live-transcribe is the streaming alternative.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-transcribe-diarize","provider":"OpenAI","model":"gpt-4o-transcribe-diarize","status":"deprecated","announcedDate":"2026-08-26","deprecationDate":"2026-08-26","sunsetDate":"2027-02-26","replacement":"gpt-transcribe","notes":"gpt-live-transcribe is the streaming alternative.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-realtime","provider":"OpenAI","model":"gpt-realtime","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-realtime-2.1","notes":"Family-level notice covering every gpt-realtime snapshot.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-realtime","provider":"OpenAI","model":"gpt-4o-realtime","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-realtime-2.1","notes":"Family-level notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-realtime-mini","provider":"OpenAI","model":"gpt-realtime-mini","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-realtime-2.1-mini","notes":"Family-level notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-mini-realtime","provider":"OpenAI","model":"gpt-4o-mini-realtime","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-realtime-2.1-mini","notes":"Family-level notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-audio","provider":"OpenAI","model":"gpt-audio","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-audio-1.5","notes":"Family-level notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-audio-mini","provider":"OpenAI","model":"gpt-audio-mini","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-audio-1.5","notes":"Family-level notice. OpenAI points mini users at the full gpt-audio-1.5, not a mini tier.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-audio","provider":"OpenAI","model":"gpt-4o-audio","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-audio-1.5","notes":"Family-level notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-mini-audio","provider":"OpenAI","model":"gpt-4o-mini-audio","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-audio-1.5","notes":"Family-level notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-mini-transcribe-2025-03-20","provider":"OpenAI","model":"gpt-4o-mini-transcribe-2025-03-20","status":"deprecated","announcedDate":"2026-07-20","deprecationDate":"2026-07-20","sunsetDate":"2027-01-20","replacement":"gpt-4o-mini-transcribe-2025-12-15","notes":"The named replacement belongs to the gpt-4o-mini-transcribe family, which itself retires on 2027-02-26.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"bedrock-claude-opus-4-1","provider":"AWS Bedrock","model":"anthropic.claude-opus-4-1-20250805-v1:0","modelDisplay":"Claude Opus 4.1 (Bedrock)","status":"deprecated","announcedDate":"2026-07-08","deprecationDate":"2026-07-08","sunsetDate":"2027-01-08","notes":"Extended-access pricing from 2026-10-08. Already retired on the Claude API (2026-08-05).","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"openai-gpt-5-2025-08-07","provider":"OpenAI","model":"gpt-5-2025-08-07","status":"deprecated","announcedDate":"2026-06-11","deprecationDate":"2026-06-11","sunsetDate":"2026-12-11","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-mini-2025-08-07","provider":"OpenAI","model":"gpt-5-mini-2025-08-07","status":"deprecated","announcedDate":"2026-06-11","deprecationDate":"2026-06-11","sunsetDate":"2026-12-11","replacement":"gpt-5.6-terra","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-nano-2025-08-07","provider":"OpenAI","model":"gpt-5-nano-2025-08-07","status":"deprecated","announcedDate":"2026-06-11","deprecationDate":"2026-06-11","sunsetDate":"2026-12-11","replacement":"gpt-5.6-luna","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-pro-2025-10-06","provider":"OpenAI","model":"gpt-5-pro-2025-10-06","status":"deprecated","announcedDate":"2026-06-11","deprecationDate":"2026-06-11","sunsetDate":"2026-12-11","replacement":"gpt-5.6-sol","notes":"The replacement is gpt-5.6-sol with reasoning.mode set to pro.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-o3-2025-04-16","provider":"OpenAI","model":"o3-2025-04-16","status":"deprecated","announcedDate":"2026-06-11","deprecationDate":"2026-06-11","sunsetDate":"2026-12-11","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-o3-pro-2025-06-10","provider":"OpenAI","model":"o3-pro-2025-06-10","status":"deprecated","announcedDate":"2026-06-11","deprecationDate":"2026-06-11","sunsetDate":"2026-12-11","replacement":"gpt-5.6-sol","notes":"The replacement is gpt-5.6-sol with reasoning.mode set to pro.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-image-1-mini","provider":"OpenAI","model":"gpt-image-1-mini","status":"deprecated","announcedDate":"2026-06-02","deprecationDate":"2026-06-02","sunsetDate":"2026-12-01","replacement":"gpt-image-2","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-image-1-5","provider":"OpenAI","model":"gpt-image-1.5","status":"deprecated","announcedDate":"2026-06-02","deprecationDate":"2026-06-02","sunsetDate":"2026-12-01","replacement":"gpt-image-2","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-chatgpt-image-latest","provider":"OpenAI","model":"chatgpt-image-latest","status":"deprecated","announcedDate":"2026-06-02","deprecationDate":"2026-06-02","sunsetDate":"2026-12-01","replacement":"gpt-image-2","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"anthropic-claude-sonnet-4-5","provider":"Anthropic","model":"claude-sonnet-4-5-20250929","modelDisplay":"Claude Sonnet 4.5","status":"deprecated","announcedDate":"2026-09-30","deprecationDate":"2026-09-30","sunsetDate":"2026-11-30","replacement":"claude-sonnet-5-5","notes":"Anthropic notified developers on September 30, 2026 and retires the model on November 30, 2026. The dates apply to Anthropic-operated platforms (Claude API, Claude Platform on AWS, Microsoft Foundry); Amazon Bedrock and Google Cloud set their own schedules.","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"bedrock-marengo-embed-2-7","provider":"AWS Bedrock","model":"twelvelabs.marengo-embed-2-7-v1:0","modelDisplay":"TwelveLabs Marengo Embed 2.7 (Bedrock)","status":"deprecated","announcedDate":"2026-05-29","deprecationDate":"2026-05-29","sunsetDate":"2026-11-30","notes":"Video embedding model. Extended-access pricing from 2026-08-29.","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"bedrock-jamba-1-5-large","provider":"AWS Bedrock","model":"ai21.jamba-1-5-large-v1:0","modelDisplay":"Jamba 1.5 Large (Bedrock)","status":"deprecated","announcedDate":"2026-05-26","deprecationDate":"2026-05-26","sunsetDate":"2026-11-26","notes":"Extended-access pricing from 2026-08-26.","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"bedrock-jamba-1-5-mini","provider":"AWS Bedrock","model":"ai21.jamba-1-5-mini-v1:0","modelDisplay":"Jamba 1.5 Mini (Bedrock)","status":"deprecated","announcedDate":"2026-05-26","deprecationDate":"2026-05-26","sunsetDate":"2026-11-26","notes":"Extended-access pricing from 2026-08-26.","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"anthropic-claude-opus-4-5","provider":"Anthropic","model":"claude-opus-4-5-20251101","modelDisplay":"Claude Opus 4.5","status":"announced","sunsetDate":"2026-11-24","notes":"Tentative retirement date on an Active model, listed as not sooner than November 24, 2026. Microsoft Foundry lists 2026-10-19 for its hosted copy, which is earlier than the Claude API date.","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"azure-o1","provider":"Microsoft Azure","model":"o1","modelDisplay":"o1 2024-12-17 (Azure)","status":"deprecated","sunsetDate":"2026-11-19","replacement":"gpt-5.6-sol","notes":"Azure OpenAI date, which runs later than the OpenAI API shutdown. Deprecated on Azure means existing subscriptions keep deploying and new ones are blocked. Provisioned deployments are not auto-upgraded.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-o1-pro","provider":"Microsoft Azure","model":"o1-pro","modelDisplay":"o1-pro 2025-03-19 (Azure)","status":"deprecated","sunsetDate":"2026-11-19","replacement":"gpt-5.6-sol","notes":"Azure OpenAI date, which runs later than the OpenAI API shutdown. Deprecated on Azure means existing subscriptions keep deploying and new ones are blocked. Provisioned deployments are not auto-upgraded.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-o3","provider":"Microsoft Azure","model":"o3","modelDisplay":"o3 2025-04-16 (Azure)","status":"deprecated","sunsetDate":"2026-11-19","replacement":"gpt-5.6-sol","notes":"Azure OpenAI date, which runs later than the OpenAI API shutdown. Deprecated on Azure means existing subscriptions keep deploying and new ones are blocked. Provisioned deployments are not auto-upgraded.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-o3-deep-research","provider":"Microsoft Azure","model":"o3-deep-research","modelDisplay":"o3-deep-research 2025-06-26 (Azure)","status":"deprecated","sunsetDate":"2026-11-19","replacement":"gpt-5.6-sol","notes":"Azure OpenAI date, which runs later than the OpenAI API shutdown. Deprecated on Azure means existing subscriptions keep deploying and new ones are blocked. Provisioned deployments are not auto-upgraded.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-o3-mini","provider":"Microsoft Azure","model":"o3-mini","modelDisplay":"o3-mini 2025-01-31 (Azure)","status":"deprecated","sunsetDate":"2026-11-19","replacement":"gpt-5.6-terra","notes":"Azure OpenAI date, which runs later than the OpenAI API shutdown. Deprecated on Azure means existing subscriptions keep deploying and new ones are blocked. Provisioned deployments are not auto-upgraded.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-o3-pro","provider":"Microsoft Azure","model":"o3-pro","modelDisplay":"o3-pro 2025-06-10 (Azure)","status":"deprecated","sunsetDate":"2026-11-19","replacement":"gpt-5.6-sol","notes":"Azure OpenAI date, which runs later than the OpenAI API shutdown. Deprecated on Azure means existing subscriptions keep deploying and new ones are blocked. Provisioned deployments are not auto-upgraded.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-o4-mini","provider":"Microsoft Azure","model":"o4-mini","modelDisplay":"o4-mini 2025-04-16 (Azure)","status":"deprecated","sunsetDate":"2026-11-19","replacement":"gpt-5.6-terra","notes":"Azure OpenAI date, which runs later than the OpenAI API shutdown. Deprecated on Azure means existing subscriptions keep deploying and new ones are blocked. Provisioned deployments are not auto-upgraded.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-codex-mini","provider":"Microsoft Azure","model":"codex-mini","modelDisplay":"codex-mini 2025-05-16 (Azure)","status":"deprecated","sunsetDate":"2026-11-15","notes":"No replacement named. The OpenAI API retired codex-mini-latest on 2026-02-12.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"xai-grok-imagine-image-quality","provider":"xAI","model":"grok-imagine-image-quality","status":"deprecated","announcedDate":"2026-09-02","deprecationDate":"2026-09-02","sunsetDate":"2026-11-02","replacement":"grok-imagine-image-2.0","notes":"After retirement the slug is served by grok-imagine-image-2.0 at low quality, one cent per image cheaper. grok-imagine-image (1.0) is not affected.","sourceUrl":"https://docs.x.ai/developers/migration/imagine-image-quality-nov-2"}
+{"id":"openai-gpt-3-5-turbo","provider":"OpenAI","model":"gpt-3.5-turbo","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-terra","notes":"Covers the alias, gpt-3.5-turbo-0125, and gpt-3.5-turbo-completions. Fine-tuned ft-gpt-3.5-turbo models retire the same day.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4","provider":"OpenAI","model":"gpt-4","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-sol","notes":"Covers the alias, gpt-4-0613, and both completions variants. Fine-tuned ft-gpt-4 models retire the same day.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4-turbo","provider":"OpenAI","model":"gpt-4-turbo","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-sol","notes":"Covers gpt-4-turbo-2024-04-09 and gpt-4-turbo-completions.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4-1-nano","provider":"OpenAI","model":"gpt-4.1-nano","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-luna","notes":"Covers gpt-4.1-nano-2025-04-14 and fine-tunes built on it. gpt-4.1 and gpt-4.1-mini are not on this notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-2024-05-13","provider":"OpenAI","model":"gpt-4o-2024-05-13","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-sol","notes":"Only the original May 2024 snapshot. The gpt-4o alias is not on the notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-image-1","provider":"OpenAI","model":"gpt-image-1","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-image-2","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-o1","provider":"OpenAI","model":"o1","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-sol","notes":"Covers the alias and o1-2024-12-17.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-o1-pro","provider":"OpenAI","model":"o1-pro","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-sol","notes":"Covers o1-pro-2025-03-19. The replacement is gpt-5.6-sol with reasoning.mode set to pro.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-o3-mini","provider":"OpenAI","model":"o3-mini","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-sol","notes":"Covers the alias and o3-mini-2025-01-31.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-o4-mini","provider":"OpenAI","model":"o4-mini","status":"deprecated","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-10-23","replacement":"gpt-5.6-terra","notes":"Covers o4-mini-2025-04-16 and fine-tuned ft-o4-mini-2025-04-16 models.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"vertex-gemini-2-5-pro","provider":"Google Vertex AI","model":"gemini-2.5-pro","modelDisplay":"Gemini 2.5 Pro (Vertex AI)","status":"announced","sunsetDate":"2026-10-20","replacement":"gemini-3.5-flash","notes":"Google Cloud retirement date. The Gemini API lists no shutdown date for this model, so the date binds Vertex AI deployments only. Google names Gemini 3.5 Flash as the replacement. Google is inconsistent here: the Vertex AI generative AI release notes entry of 2026-04-02 gives 2026-10-16 for the whole Gemini 2.5 family, four days earlier, so plan against the earlier date. Separately, on 2026-09-18 the Gemini API limited 2.5 access to users who have used the models before, while stating they are not deprecated; Google points new projects to Gemini 3.5 Flash-Lite or 3.8 Flash.","sourceUrl":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/model-versions"}
+{"id":"vertex-gemini-2-5-flash","provider":"Google Vertex AI","model":"gemini-2.5-flash","modelDisplay":"Gemini 2.5 Flash (Vertex AI)","status":"announced","sunsetDate":"2026-10-20","replacement":"gemini-3.5-flash-lite","notes":"Google Cloud retirement date; the Gemini API lists none. Google names Gemini 3.5 Flash-Lite or Gemini 3.1 Flash-Lite. The Vertex AI generative AI release notes give 2026-10-16 for the same model, so plan against the earlier date. Separately, on 2026-09-18 the Gemini API limited 2.5 access to users who have used the models before, while stating they are not deprecated; Google points new projects to Gemini 3.5 Flash-Lite or 3.8 Flash.","sourceUrl":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/model-versions"}
+{"id":"vertex-gemini-2-5-flash-lite","provider":"Google Vertex AI","model":"gemini-2.5-flash-lite","modelDisplay":"Gemini 2.5 Flash-Lite (Vertex AI)","status":"announced","sunsetDate":"2026-10-20","replacement":"gemini-3.1-flash-lite","notes":"Google Cloud retirement date; the Gemini API lists none. Google names Gemini 3.1 Flash-Lite or Gemma 4. The Vertex AI generative AI release notes give 2026-10-16 for the same model, so plan against the earlier date. Separately, on 2026-09-18 the Gemini API limited 2.5 access to users who have used the models before, while stating they are not deprecated; Google points new projects to Gemini 3.5 Flash-Lite or 3.8 Flash.","sourceUrl":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/model-versions"}
+{"id":"anthropic-claude-haiku-4-5","provider":"Anthropic","model":"claude-haiku-4-5-20251001","modelDisplay":"Claude Haiku 4.5","status":"announced","sunsetDate":"2026-10-15","notes":"Tentative retirement date on an Active model, listed as not sooner than October 15, 2026. Microsoft Foundry lists 2026-10-19 for both of its hosted versions. This is the model Anthropic names as the replacement on every retired Haiku row above.","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"bedrock-claude-sonnet-4","provider":"AWS Bedrock","model":"anthropic.claude-sonnet-4-20250514-v1:0","modelDisplay":"Claude Sonnet 4 (Bedrock)","status":"deprecated","announcedDate":"2026-04-14","deprecationDate":"2026-04-14","sunsetDate":"2026-10-14","notes":"In public extended access since 2026-07-14, where Anthropic can set higher prices. Already retired on the Claude API.","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"google-antigravity-preview-05-2026","provider":"Google","model":"antigravity-preview-05-2026","status":"announced","sunsetDate":"2026-10-05","replacement":"antigravity-preview-09-2026","notes":"Managed agent rather than a model, listed in the Managed Agents section of the Gemini API deprecations page. The September 2026 preview that replaces it shipped on 2026-09-17, leaving about three weeks of overlap.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-gemini-2-5-flash-image","provider":"Google","model":"gemini-2.5-flash-image","status":"announced","sunsetDate":"2026-10-02","replacement":"gemini-3.1-flash-lite-image","notes":"Same date on the Gemini API and Vertex AI. The Gemini API table points to gemini-3.1-flash-image-preview, which is itself deprecated; Vertex AI names Gemini 3.1 Flash-Lite Image.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"openai-gpt-5-4-cyber","provider":"OpenAI","model":"gpt-5.4-cyber","status":"deprecated","announcedDate":"2026-09-11","deprecationDate":"2026-09-11","sunsetDate":"2026-10-01","replacement":"gpt-5.6-cyber","notes":"Announced with 20 days of notice.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"azure-gpt-4o-2024-05-13","provider":"Microsoft Azure","model":"gpt-4o-2024-05-13","modelDisplay":"gpt-4o 2024-05-13 (Azure)","status":"deprecated","sunsetDate":"2026-10-01","replacement":"gpt-5.1","notes":"Standard deployments auto-upgrade to gpt-5.1 at retirement; provisioned deployments do not and must migrate by hand. About three weeks ahead of the OpenAI API shutdown.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"google-gemini-omni-flash-preview","provider":"Google","model":"gemini-omni-flash-preview","status":"deprecated","announcedDate":"2026-08-27","deprecationDate":"2026-08-27","sunsetDate":"2026-09-30","replacement":"gemini-omni-1.1-flash","notes":"Video generation preview that lasted three months before the GA release replaced it.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"bedrock-nova-canvas","provider":"AWS Bedrock","model":"amazon.nova-canvas-v1:0","modelDisplay":"Nova Canvas (Bedrock)","status":"deprecated","announcedDate":"2026-03-30","deprecationDate":"2026-03-30","sunsetDate":"2026-09-30","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"bedrock-nova-reel","provider":"AWS Bedrock","model":"amazon.nova-reel-v1:0","modelDisplay":"Nova Reel (Bedrock)","status":"deprecated","announcedDate":"2026-03-30","deprecationDate":"2026-03-30","sunsetDate":"2026-09-30","notes":"amazon.nova-reel-v1:1 reaches end of life the same day.","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"openai-gpt-3-5-turbo-instruct","provider":"OpenAI","model":"gpt-3.5-turbo-instruct","status":"deprecated","announcedDate":"2025-09-26","deprecationDate":"2025-09-26","sunsetDate":"2026-09-28","replacement":"gpt-5.6-terra","notes":"The last instruct-style completion model on the API.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-babbage-002","provider":"OpenAI","model":"babbage-002","status":"deprecated","announcedDate":"2025-09-26","deprecationDate":"2025-09-26","sunsetDate":"2026-09-28","replacement":"gpt-5.6-terra","notes":"Base model shuts down 2026-09-28. Fine-tuned ft-babbage-002 models are on a separate notice and keep serving until 2026-10-23.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-davinci-002","provider":"OpenAI","model":"davinci-002","status":"deprecated","announcedDate":"2025-09-26","deprecationDate":"2025-09-26","sunsetDate":"2026-09-28","replacement":"gpt-5.6-terra","notes":"Base model shuts down 2026-09-28. Fine-tuned ft-davinci-002 models are on a separate notice and keep serving until 2026-10-23.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-3-5-turbo-1106","provider":"OpenAI","model":"gpt-3.5-turbo-1106","status":"deprecated","announcedDate":"2025-09-26","deprecationDate":"2025-09-26","sunsetDate":"2026-09-28","replacement":"gpt-5.6-terra","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-sora-2","provider":"OpenAI","model":"sora-2","status":"deprecated","announcedDate":"2026-03-24","deprecationDate":"2026-03-24","sunsetDate":"2026-09-24","notes":"The whole Videos API shuts down the same day and OpenAI names no replacement. Snapshots sora-2-2025-10-06 and sora-2-2025-12-08 go with it.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-sora-2-pro","provider":"OpenAI","model":"sora-2-pro","status":"deprecated","announcedDate":"2026-03-24","deprecationDate":"2026-03-24","sunsetDate":"2026-09-24","notes":"Includes sora-2-pro-2025-10-06. No replacement named; the Videos API itself is retired.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"bedrock-nova-premier","provider":"AWS Bedrock","model":"amazon.nova-premier-v1:0","modelDisplay":"Nova Premier (Bedrock)","status":"sunsetted","announcedDate":"2026-03-13","deprecationDate":"2026-03-13","sunsetDate":"2026-09-14","notes":"End of life on 2026-09-14. The lifecycle table lists no extended-access start date.","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"bedrock-nova-sonic","provider":"AWS Bedrock","model":"amazon.nova-sonic-v1:0","modelDisplay":"Nova Sonic (Bedrock)","status":"sunsetted","announcedDate":"2026-03-13","deprecationDate":"2026-03-13","sunsetDate":"2026-09-14","notes":"End of life on 2026-09-14. The lifecycle table lists no extended-access start date.","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"bedrock-claude-3-haiku","provider":"AWS Bedrock","model":"anthropic.claude-3-haiku-20240307-v1:0","modelDisplay":"Claude 3 Haiku (Bedrock)","status":"sunsetted","announcedDate":"2026-03-10","deprecationDate":"2026-03-10","sunsetDate":"2026-09-10","notes":"End of life in commercial regions and GovCloud. Extended-access pricing applied from 2026-06-10.","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"deepseek-v4-flash","provider":"DeepSeek","model":"deepseek-v4-flash","modelDisplay":"DeepSeek V4 Flash","status":"sunsetted","announcedDate":"2026-09-10","deprecationDate":"2026-09-10","sunsetDate":"2026-09-10","replacement":"deepseek-flash","notes":"V4 Flash was retired when V4.1 Flash launched. The deepseek-v4-flash name is temporarily routed to V4.1 Flash, so calls still succeed but hit a different model. deepseek-flash is the new name for the current Flash model. V4 Pro is unaffected and DeepSeek says it will keep serving past 2026-09-14.","sourceUrl":"https://api-docs.deepseek.com/updates"}
+{"id":"deepseek-v4-flash-vision-exp","provider":"DeepSeek","model":"deepseek-v4-flash-vision-exp","status":"sunsetted","announcedDate":"2026-09-10","deprecationDate":"2026-09-10","sunsetDate":"2026-09-10","replacement":"deepseek-flash","notes":"Experimental vision model launched 2026-08-21 and retired three weeks later. The name temporarily routes to V4.1 Flash, which has native vision.","sourceUrl":"https://api-docs.deepseek.com/updates"}
+{"id":"google-gemini-robotics-er-1-6-preview","provider":"Google","model":"gemini-robotics-er-1.6-preview","status":"deprecated","announcedDate":"2026-07-30","deprecationDate":"2026-07-30","sunsetDate":"2026-08-31","replacement":"gemini-robotics-er-2-preview","notes":"Past the scheduled date but not yet marked shut down.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"mistral-medium-2505","provider":"Mistral","model":"mistral-medium-2505","modelDisplay":"Mistral Medium 3","status":"sunsetted","deprecationDate":"2026-05-22","sunsetDate":"2026-08-31","replacement":"mistral-medium-3-5","sourceUrl":"https://docs.mistral.ai/models"}
+{"id":"mistral-medium-2508","provider":"Mistral","model":"mistral-medium-2508","modelDisplay":"Mistral Medium 3.1","status":"sunsetted","deprecationDate":"2026-05-22","sunsetDate":"2026-08-31","replacement":"mistral-medium-3-5","sourceUrl":"https://docs.mistral.ai/models"}
+{"id":"vertex-claude-3-haiku","provider":"Google Vertex AI","model":"claude-3-haiku","modelDisplay":"Claude 3 Haiku (Google Cloud)","status":"sunsetted","announcedDate":"2026-02-23","deprecationDate":"2026-02-23","sunsetDate":"2026-08-23","notes":"Google Cloud copy. The Claude API retired it on 2026-04-20.","sourceUrl":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/deprecations/partner-models"}
+{"id":"bedrock-command-r","provider":"AWS Bedrock","model":"cohere.command-r-v1:0","modelDisplay":"Command R (Bedrock)","status":"sunsetted","announcedDate":"2026-02-19","deprecationDate":"2026-02-19","sunsetDate":"2026-08-19","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"bedrock-command-r-plus","provider":"AWS Bedrock","model":"cohere.command-r-plus-v1:0","modelDisplay":"Command R+ (Bedrock)","status":"sunsetted","announcedDate":"2026-02-19","deprecationDate":"2026-02-19","sunsetDate":"2026-08-19","sourceUrl":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle-legacy.html"}
+{"id":"google-imagen-4-0-generate-001","provider":"Google","model":"imagen-4.0-generate-001","status":"deprecated","announcedDate":"2026-06-15","deprecationDate":"2026-06-15","sunsetDate":"2026-08-17","replacement":"gemini-3.1-flash-image","notes":"Google is folding image generation into the Gemini image models. Past the scheduled date but not yet marked shut down.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-imagen-4-0-ultra-generate-001","provider":"Google","model":"imagen-4.0-ultra-generate-001","status":"deprecated","announcedDate":"2026-06-15","deprecationDate":"2026-06-15","sunsetDate":"2026-08-17","replacement":"gemini-3.1-flash-image","notes":"Past the scheduled date but not yet marked shut down.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-imagen-4-0-fast-generate-001","provider":"Google","model":"imagen-4.0-fast-generate-001","status":"deprecated","announcedDate":"2026-06-15","deprecationDate":"2026-06-15","sunsetDate":"2026-08-17","replacement":"gemini-3.1-flash-image","notes":"Past the scheduled date but not yet marked shut down.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"azure-mai-image-2","provider":"Microsoft Azure","model":"MAI-Image-2","status":"sunsetted","sunsetDate":"2026-08-15","replacement":"MAI-Image-2.5","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-mai-image-2e","provider":"Microsoft Azure","model":"MAI-Image-2e","status":"sunsetted","sunsetDate":"2026-08-15","replacement":"MAI-Image-2.5-Flash","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-deepseek-r1","provider":"Microsoft Azure","model":"DeepSeek-R1","modelDisplay":"DeepSeek-R1 (Azure)","status":"sunsetted","sunsetDate":"2026-08-13","replacement":"DeepSeek-V4-Pro","notes":"Outlived DeepSeek-R1-0528 on Foundry by one month.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"openai-gpt-5-2-chat-latest","provider":"OpenAI","model":"gpt-5.2-chat-latest","status":"sunsetted","announcedDate":"2026-05-08","deprecationDate":"2026-05-08","sunsetDate":"2026-08-10","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-3-chat-latest","provider":"OpenAI","model":"gpt-5.3-chat-latest","status":"sunsetted","announcedDate":"2026-05-08","deprecationDate":"2026-05-08","sunsetDate":"2026-08-10","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"google-gemini-embedding-2-preview","provider":"Google","model":"gemini-embedding-2-preview","status":"sunsetted","sunsetDate":"2026-08-10","replacement":"gemini-embedding-2","notes":"Multimodal embedding preview released 2026-03-10. The deprecations table lists it as embedding-2-preview. Vectors from a preview model should be re-embedded, not mixed with GA output.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"anthropic-claude-opus-4-1","provider":"Anthropic","model":"claude-opus-4-1-20250805","modelDisplay":"Claude Opus 4.1","status":"sunsetted","announcedDate":"2026-06-05","deprecationDate":"2026-06-05","sunsetDate":"2026-08-05","replacement":"claude-opus-4-8","notes":"Retired on the Claude API and Microsoft Foundry on 2026-08-05 (Foundry names claude-opus-5 as its replacement). Bedrock keeps it in extended access until 2027-01-08.","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"mistral-small-2506","provider":"Mistral","model":"mistral-small-2506","modelDisplay":"Mistral Small 3.2","status":"sunsetted","deprecationDate":"2026-04-30","sunsetDate":"2026-07-31","replacement":"mistral-small-2603","sourceUrl":"https://docs.mistral.ai/models"}
+{"id":"mistral-magistral-small-2509","provider":"Mistral","model":"magistral-small-2509","modelDisplay":"Magistral Small 1.2","status":"sunsetted","deprecationDate":"2026-04-30","sunsetDate":"2026-07-31","replacement":"mistral-small-2603","notes":"Mistral Small 4 unifies instruct, reasoning, and coding, so reasoning workloads move there.","sourceUrl":"https://docs.mistral.ai/models"}
+{"id":"mistral-magistral-medium-2509","provider":"Mistral","model":"magistral-medium-2509","modelDisplay":"Magistral Medium 1.2","status":"sunsetted","deprecationDate":"2026-05-22","sunsetDate":"2026-07-31","replacement":"mistral-medium-3-5","sourceUrl":"https://docs.mistral.ai/models"}
+{"id":"mistral-devstral-2512","provider":"Mistral","model":"devstral-2512","modelDisplay":"Devstral 2","status":"sunsetted","deprecationDate":"2026-05-22","sunsetDate":"2026-07-31","replacement":"mistral-medium-3-5","notes":"Coding agents move to the general Medium 3.5 model rather than a dedicated Devstral successor.","sourceUrl":"https://docs.mistral.ai/models"}
+{"id":"mistral-open-mistral-nemo-2407","provider":"Mistral","model":"open-mistral-nemo-2407","modelDisplay":"Mistral Nemo 12B","status":"sunsetted","deprecationDate":"2026-05-22","sunsetDate":"2026-07-31","replacement":"ministral-8b-2512","sourceUrl":"https://docs.mistral.ai/models"}
+{"id":"deepseek-chat-alias","provider":"DeepSeek","model":"deepseek-chat","status":"sunsetted","announcedDate":"2026-07-14","deprecationDate":"2026-07-24","sunsetDate":"2026-07-24","replacement":"deepseek-v4-pro","notes":"Legacy alias stopped serving on 2026-07-24 at 15:59 UTC. Same base URL and key, so migration is a one-line model parameter change.","sourceUrl":"https://api-docs.deepseek.com/updates"}
+{"id":"deepseek-reasoner-alias","provider":"DeepSeek","model":"deepseek-reasoner","status":"sunsetted","announcedDate":"2026-07-14","deprecationDate":"2026-07-24","sunsetDate":"2026-07-24","replacement":"deepseek-v4-flash","notes":"Legacy alias stopped serving on 2026-07-24 at 15:59 UTC. Mapping trap: deepseek-reasoner routed to V4 Flash in thinking mode, not to the stronger V4 Pro, so a naive swap to deepseek-v4-pro changes cost. Teams running heavy reasoning on the old alias should pick the tier deliberately.","sourceUrl":"https://api-docs.deepseek.com/updates"}
+{"id":"openai-computer-use-preview","provider":"OpenAI","model":"computer-use-preview","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-terra","notes":"The alias and computer-use-preview-2025-03-11 were both removed.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-search-preview","provider":"OpenAI","model":"gpt-4o-search-preview-2025-03-11","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-terra","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-4o-mini-search-preview","provider":"OpenAI","model":"gpt-4o-mini-search-preview-2025-03-11","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-terra","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-chat-latest","provider":"OpenAI","model":"gpt-5-chat-latest","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-codex","provider":"OpenAI","model":"gpt-5-codex","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-1-chat-latest","provider":"OpenAI","model":"gpt-5.1-chat-latest","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-1-codex","provider":"OpenAI","model":"gpt-5.1-codex","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-1-codex-max","provider":"OpenAI","model":"gpt-5.1-codex-max","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-1-codex-mini","provider":"OpenAI","model":"gpt-5.1-codex-mini","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-terra","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-5-2-codex","provider":"OpenAI","model":"gpt-5.2-codex","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-sol","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-audio-mini-2025-10-06","provider":"OpenAI","model":"gpt-audio-mini-2025-10-06","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-audio-1.5","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-realtime-mini-2025-10-06","provider":"OpenAI","model":"gpt-realtime-mini-2025-10-06","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-realtime-2.1-mini","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-o3-deep-research","provider":"OpenAI","model":"o3-deep-research","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-sol","notes":"The alias and o3-deep-research-2025-06-26 were both removed. Azure keeps its copy until 2026-11-19.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-o4-mini-deep-research","provider":"OpenAI","model":"o4-mini-deep-research","status":"sunsetted","announcedDate":"2026-04-22","deprecationDate":"2026-04-22","sunsetDate":"2026-07-23","replacement":"gpt-5.6-sol","notes":"The alias and o4-mini-deep-research-2025-06-26 were both removed.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"azure-mistral-document-ai-2505","provider":"Microsoft Azure","model":"mistral-document-ai-2505","status":"sunsetted","sunsetDate":"2026-07-20","replacement":"mistral-document-ai-2512","notes":"Azure also names mistral-ocr-4-0 as a replacement.","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-deepseek-r1-0528","provider":"Microsoft Azure","model":"DeepSeek-R1-0528","modelDisplay":"DeepSeek-R1-0528 (Azure)","status":"sunsetted","sunsetDate":"2026-07-13","replacement":"DeepSeek-V4-Pro","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-deepseek-v3-0324","provider":"Microsoft Azure","model":"DeepSeek-V3-0324","modelDisplay":"DeepSeek-V3-0324 (Azure)","status":"sunsetted","sunsetDate":"2026-07-13","replacement":"DeepSeek-V4-Flash","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"azure-deepseek-v3-1","provider":"Microsoft Azure","model":"DeepSeek-V3.1","modelDisplay":"DeepSeek-V3.1 (Azure)","status":"sunsetted","sunsetDate":"2026-07-13","replacement":"DeepSeek-V4-Flash","sourceUrl":"https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule"}
+{"id":"vertex-claude-3-5-haiku","provider":"Google Vertex AI","model":"claude-3-5-haiku","modelDisplay":"Claude 3.5 Haiku (Google Cloud)","status":"sunsetted","announcedDate":"2026-01-05","deprecationDate":"2026-01-05","sunsetDate":"2026-07-05","notes":"Google Cloud copy, available to existing customers only during deprecation. The Claude API retired it on 2026-02-19.","sourceUrl":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/deprecations/partner-models"}
+{"id":"google-veo-2-0-generate-001","provider":"Google","model":"veo-2.0-generate-001","status":"sunsetted","announcedDate":"2026-06-15","deprecationDate":"2026-06-15","sunsetDate":"2026-06-30","replacement":"veo-3.1-generate-preview","notes":"Fifteen days of notice. GA Veo 3.1 models are available on Google Cloud.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-veo-3-0-generate-001","provider":"Google","model":"veo-3.0-generate-001","status":"deprecated","announcedDate":"2026-06-15","deprecationDate":"2026-06-15","sunsetDate":"2026-06-30","replacement":"veo-3.1-generate-preview","notes":"Past its 2026-06-30 date on both the Gemini API and Vertex AI but not yet marked shut down. Vertex AI names veo-3.1-generate-001 as the GA replacement.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-veo-3-0-fast-generate-001","provider":"Google","model":"veo-3.0-fast-generate-001","status":"deprecated","announcedDate":"2026-06-15","deprecationDate":"2026-06-15","sunsetDate":"2026-06-30","replacement":"veo-3.1-fast-generate-preview","notes":"Past its 2026-06-30 date on both the Gemini API and Vertex AI but not yet marked shut down. Vertex AI names veo-3.1-fast-generate-001 as the GA replacement.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-gemini-3-1-flash-image-preview","provider":"Google","model":"gemini-3.1-flash-image-preview","status":"deprecated","announcedDate":"2026-05-28","deprecationDate":"2026-05-28","sunsetDate":"2026-06-25","replacement":"gemini-3.1-flash-image","notes":"Past its scheduled shutdown but not yet marked shut down on the deprecations table. Plan as if it is gone.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-gemini-3-pro-image-preview","provider":"Google","model":"gemini-3-pro-image-preview","status":"deprecated","announcedDate":"2026-05-28","deprecationDate":"2026-05-28","sunsetDate":"2026-06-25","replacement":"gemini-3-pro-image","notes":"Past its scheduled shutdown but not yet marked shut down on the deprecations table. Plan as if it is gone.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"anthropic-claude-sonnet-4","provider":"Anthropic","model":"claude-sonnet-4-20250514","modelDisplay":"Claude Sonnet 4","status":"sunsetted","announcedDate":"2026-04-14","deprecationDate":"2026-04-14","sunsetDate":"2026-06-15","replacement":"claude-sonnet-4-6","notes":"Claude API date. Bedrock keeps it in extended access until 2026-10-14.","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"anthropic-claude-opus-4","provider":"Anthropic","model":"claude-opus-4-20250514","modelDisplay":"Claude Opus 4","status":"sunsetted","announcedDate":"2026-04-14","deprecationDate":"2026-04-14","sunsetDate":"2026-06-15","replacement":"claude-opus-4-8","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"anthropic-claude-mythos-preview","provider":"Anthropic","model":"claude-mythos-preview","modelDisplay":"Claude Mythos Preview","status":"deprecated","announcedDate":"2026-06-09","deprecationDate":"2026-06-09","replacement":"claude-mythos-5","notes":"Invitation-only research preview. Anthropic lists it as deprecated but has not published a retirement date; Microsoft Foundry shows 2027-04-02 for its hosted copy. claude-fable-5 is the same capability without access gating.","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"google-gemini-2-0-flash","provider":"Google","model":"gemini-2.0-flash","modelDisplay":"Gemini 2.0 Flash","status":"sunsetted","announcedDate":"2026-02-18","deprecationDate":"2026-02-18","sunsetDate":"2026-06-01","replacement":"gemini-3.6-flash","notes":"Shut down on the Gemini API and retired on Vertex AI the same day, including gemini-2.0-flash-001. The Gemini API table names gemini-3.6-flash; Vertex AI names gemini-3.1-flash-lite.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"google-gemini-2-0-flash-lite","provider":"Google","model":"gemini-2.0-flash-lite","modelDisplay":"Gemini 2.0 Flash-Lite","status":"sunsetted","announcedDate":"2026-02-18","deprecationDate":"2026-02-18","sunsetDate":"2026-06-01","replacement":"gemini-3.1-flash-lite","notes":"Includes gemini-2.0-flash-lite-001. Retired on Vertex AI the same day.","sourceUrl":"https://ai.google.dev/gemini-api/docs/deprecations"}
+{"id":"xai-grok-4-1-fast-reasoning","provider":"xAI","model":"grok-4-1-fast-reasoning","status":"sunsetted","sunsetDate":"2026-05-15","replacement":"grok-4.3","notes":"Slug now served by grok-4.3 at low reasoning effort and billed at grok-4.3 rates.","sourceUrl":"https://docs.x.ai/developers/migration/may-15-retirement"}
+{"id":"xai-grok-4-1-fast-non-reasoning","provider":"xAI","model":"grok-4-1-fast-non-reasoning","status":"sunsetted","sunsetDate":"2026-05-15","replacement":"grok-4.3","notes":"Slug now served by grok-4.3 with reasoning effort none and billed at grok-4.3 rates.","sourceUrl":"https://docs.x.ai/developers/migration/may-15-retirement"}
+{"id":"xai-grok-4-fast-reasoning","provider":"xAI","model":"grok-4-fast-reasoning","status":"sunsetted","sunsetDate":"2026-05-15","replacement":"grok-4.3","notes":"Slug now served by grok-4.3 at low reasoning effort and billed at grok-4.3 rates.","sourceUrl":"https://docs.x.ai/developers/migration/may-15-retirement"}
+{"id":"xai-grok-4-fast-non-reasoning","provider":"xAI","model":"grok-4-fast-non-reasoning","status":"sunsetted","sunsetDate":"2026-05-15","replacement":"grok-4.3","notes":"Slug now served by grok-4.3 with reasoning effort none and billed at grok-4.3 rates.","sourceUrl":"https://docs.x.ai/developers/migration/may-15-retirement"}
+{"id":"xai-grok-4-0709","provider":"xAI","model":"grok-4-0709","status":"sunsetted","sunsetDate":"2026-05-15","replacement":"grok-4.3","notes":"Slug now served by grok-4.3 at low reasoning effort. Set medium or high effort explicitly to match the old depth.","sourceUrl":"https://docs.x.ai/developers/migration/may-15-retirement"}
+{"id":"xai-grok-code-fast-1","provider":"xAI","model":"grok-code-fast-1","status":"sunsetted","sunsetDate":"2026-05-15","replacement":"grok-build-0.1","notes":"Slug now routes to grok-build-0.1.","sourceUrl":"https://docs.x.ai/developers/migration/may-15-retirement"}
+{"id":"xai-grok-3","provider":"xAI","model":"grok-3","status":"sunsetted","sunsetDate":"2026-05-15","replacement":"grok-4.3","notes":"Slug now served by grok-4.3 with reasoning effort none. Azure retired its grok-3 copy on 2026-05-01.","sourceUrl":"https://docs.x.ai/developers/migration/may-15-retirement"}
+{"id":"xai-grok-imagine-image-pro","provider":"xAI","model":"grok-imagine-image-pro","status":"sunsetted","sunsetDate":"2026-05-15","replacement":"grok-imagine-image-quality","notes":"Redirects to grok-imagine-image-quality, which is itself retired on 2026-11-02.","sourceUrl":"https://docs.x.ai/developers/migration/may-15-retirement"}
+{"id":"anthropic-claude-3-haiku","provider":"Anthropic","model":"claude-3-haiku-20240307","modelDisplay":"Claude Haiku 3","status":"sunsetted","announcedDate":"2026-02-19","deprecationDate":"2026-02-19","sunsetDate":"2026-04-20","replacement":"claude-haiku-4-5-20251001","notes":"Claude API date. The Google Cloud copy ran until 2026-08-23 and the Bedrock copy until 2026-09-10.","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"cohere-embed-english-v2","provider":"Cohere","model":"embed-english-v2.0","status":"deprecated","announcedDate":"2026-04-04","deprecationDate":"2026-04-04","replacement":"embed-v4.0","notes":"embed-english-light-v2.0 is on the same notice. Cohere also names embed-english-v3.0. No shutdown date published. Changing embedding model means re-embedding the corpus, so treat this as a project, not a config change.","sourceUrl":"https://docs.cohere.com/docs/deprecations"}
+{"id":"cohere-embed-multilingual-v2","provider":"Cohere","model":"embed-multilingual-v2.0","status":"deprecated","announcedDate":"2026-04-04","deprecationDate":"2026-04-04","replacement":"embed-v4.0","notes":"Cohere also names embed-multilingual-v3.0. No shutdown date published.","sourceUrl":"https://docs.cohere.com/docs/deprecations"}
+{"id":"cohere-c4ai-aya-expanse-8b","provider":"Cohere","model":"c4ai-aya-expanse-8b","status":"deprecated","announcedDate":"2026-04-04","deprecationDate":"2026-04-04","replacement":"command-a-03-2025","notes":"Cohere Labs multilingual model. Cohere also names command-r7b-12-2024 and command-a-reasoning-08-2025. No shutdown date published.","sourceUrl":"https://docs.cohere.com/docs/deprecations"}
+{"id":"cohere-c4ai-aya-vision-8b","provider":"Cohere","model":"c4ai-aya-vision-8b","status":"deprecated","announcedDate":"2026-04-04","deprecationDate":"2026-04-04","replacement":"command-a-03-2025","notes":"Retired alongside c4ai-aya-expanse-8b. The named replacements are text models, so vision workloads need a different target. No shutdown date published.","sourceUrl":"https://docs.cohere.com/docs/deprecations"}
+{"id":"openai-gpt-4-0314","provider":"OpenAI","model":"gpt-4-0314","status":"sunsetted","announcedDate":"2025-09-26","deprecationDate":"2025-09-26","sunsetDate":"2026-03-26","replacement":"gpt-5","notes":"Limited to existing users since 2024, then formally retired in the September 2025 legacy batch alongside gpt-4-0125-preview. OpenAI named gpt-5 or gpt-4.1 as the target; the gpt-5 launch snapshot itself retires on 2026-12-11.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"anthropic-claude-3-7-sonnet","provider":"Anthropic","model":"claude-3-7-sonnet-20250219","modelDisplay":"Claude Sonnet 3.7","status":"sunsetted","announcedDate":"2025-10-28","deprecationDate":"2025-10-28","sunsetDate":"2026-02-19","replacement":"claude-sonnet-4-6","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"anthropic-claude-3-5-haiku","provider":"Anthropic","model":"claude-3-5-haiku-20241022","modelDisplay":"Claude Haiku 3.5","status":"sunsetted","announcedDate":"2025-12-19","deprecationDate":"2025-12-19","sunsetDate":"2026-02-19","replacement":"claude-haiku-4-5-20251001","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"anthropic-claude-3-opus","provider":"Anthropic","model":"claude-3-opus-20240229","modelDisplay":"Claude Opus 3","status":"sunsetted","announcedDate":"2025-06-30","deprecationDate":"2025-06-30","sunsetDate":"2026-01-05","replacement":"claude-opus-4-8","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"anthropic-claude-3-5-sonnet-20240620","provider":"Anthropic","model":"claude-3-5-sonnet-20240620","status":"sunsetted","announcedDate":"2025-08-13","deprecationDate":"2025-08-13","sunsetDate":"2025-10-28","replacement":"claude-sonnet-4-6","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"anthropic-claude-3-5-sonnet-20241022","provider":"Anthropic","model":"claude-3-5-sonnet-20241022","status":"sunsetted","announcedDate":"2025-08-13","deprecationDate":"2025-08-13","sunsetDate":"2025-10-28","replacement":"claude-sonnet-4-6","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"cohere-command","provider":"Cohere","model":"command","modelDisplay":"Cohere Command (original)","status":"deprecated","deprecationDate":"2025-09-15","replacement":"command-a-03-2025","notes":"Deprecated in the September 2025 batch with command-light, the legacy generate and summarize endpoints, and fine-tuning. Cohere has not published a shutdown date; deprecated models stay open to existing users until one is assigned.","sourceUrl":"https://docs.cohere.com/docs/deprecations"}
+{"id":"cohere-command-light","provider":"Cohere","model":"command-light","status":"deprecated","deprecationDate":"2025-09-15","replacement":"command-a-03-2025","notes":"No shutdown date published.","sourceUrl":"https://docs.cohere.com/docs/deprecations"}
+{"id":"cohere-command-r-alias","provider":"Cohere","model":"command-r","status":"deprecated","deprecationDate":"2025-09-15","replacement":"command-a-03-2025","notes":"The command-r alias and command-r-03-2024 are deprecated. Cohere suggests command-r-08-2024 or command-a-03-2025. Bedrock retired its Command R copy on 2026-08-19.","sourceUrl":"https://docs.cohere.com/docs/deprecations"}
+{"id":"cohere-command-r-plus-alias","provider":"Cohere","model":"command-r-plus","status":"deprecated","deprecationDate":"2025-09-15","replacement":"command-a-03-2025","notes":"The command-r-plus alias and command-r-plus-04-2024 are deprecated. Cohere suggests command-r-plus-08-2024 or command-a-03-2025. Bedrock retired its Command R+ copy on 2026-08-19.","sourceUrl":"https://docs.cohere.com/docs/deprecations"}
+{"id":"anthropic-claude-2-0","provider":"Anthropic","model":"claude-2.0","status":"sunsetted","announcedDate":"2025-01-21","deprecationDate":"2025-01-21","sunsetDate":"2025-07-21","replacement":"claude-opus-4-8","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"anthropic-claude-2-1","provider":"Anthropic","model":"claude-2.1","status":"sunsetted","announcedDate":"2025-01-21","deprecationDate":"2025-01-21","sunsetDate":"2025-07-21","replacement":"claude-opus-4-8","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"anthropic-claude-3-sonnet","provider":"Anthropic","model":"claude-3-sonnet-20240229","modelDisplay":"Claude Sonnet 3","status":"sunsetted","announcedDate":"2025-01-21","deprecationDate":"2025-01-21","sunsetDate":"2025-07-21","replacement":"claude-sonnet-4-6","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"openai-gpt-4-32k","provider":"OpenAI","model":"gpt-4-32k-0314","status":"sunsetted","announcedDate":"2024-06-06","deprecationDate":"2024-06-06","sunsetDate":"2025-06-06","replacement":"gpt-4o","notes":"All three 32k variants (gpt-4-32k, gpt-4-32k-0314, gpt-4-32k-0613) shut down together.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"google-gemini-1-0-pro","provider":"Google","model":"gemini-1.0-pro","modelDisplay":"Gemini 1.0 Pro","status":"sunsetted","deprecationDate":"2025-02-15","sunsetDate":"2025-02-15","replacement":"gemini-1.5-pro","notes":"First-generation Gemini API. Replaced by 1.5 family with longer context.","sourceUrl":"https://ai.google.dev/gemini-api/docs/changelog"}
+{"id":"anthropic-claude-instant-1","provider":"Anthropic","model":"claude-instant-1.2","modelDisplay":"Claude Instant 1.2","status":"sunsetted","announcedDate":"2024-09-04","deprecationDate":"2024-09-04","sunsetDate":"2024-11-06","replacement":"claude-haiku-4-5-20251001","notes":"Original cost-tier Claude. Retired with the rest of the Claude 1 family (claude-1.0 through claude-1.3 and the earlier Instant versions).","sourceUrl":"https://platform.claude.com/docs/en/about-claude/model-deprecations"}
+{"id":"google-text-bison-001","provider":"Google","model":"text-bison-001","modelDisplay":"PaLM 2 (text-bison)","status":"sunsetted","announcedDate":"2024-04-09","deprecationDate":"2024-08-15","sunsetDate":"2024-10-09","replacement":"gemini-1.5-pro","notes":"Original PaLM 2 text generation API. Migration to Gemini API required.","sourceUrl":"https://ai.google.dev/gemini-api/docs/changelog"}
+{"id":"google-chat-bison-001","provider":"Google","model":"chat-bison-001","modelDisplay":"PaLM 2 (chat-bison)","status":"sunsetted","announcedDate":"2024-04-09","deprecationDate":"2024-08-15","sunsetDate":"2024-10-09","replacement":"gemini-1.5-pro","sourceUrl":"https://ai.google.dev/gemini-api/docs/changelog"}
+{"id":"openai-gpt-3-5-turbo-0301","provider":"OpenAI","model":"gpt-3.5-turbo-0301","status":"sunsetted","announcedDate":"2023-06-13","deprecationDate":"2023-06-13","sunsetDate":"2024-09-13","replacement":"gpt-3.5-turbo","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-gpt-3-5-turbo-0613","provider":"OpenAI","model":"gpt-3.5-turbo-0613","status":"sunsetted","announcedDate":"2023-11-06","deprecationDate":"2023-11-06","sunsetDate":"2024-09-13","replacement":"gpt-3.5-turbo","notes":"gpt-3.5-turbo-16k-0613 shut down the same day.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}
+{"id":"openai-text-davinci-003","provider":"OpenAI","model":"text-davinci-003","status":"sunsetted","announcedDate":"2023-07-06","deprecationDate":"2023-07-06","sunsetDate":"2024-01-04","replacement":"gpt-3.5-turbo-instruct","notes":"Last of the original GPT-3 completion-API family. Replaced by chat-completions.","sourceUrl":"https://developers.openai.com/api/docs/deprecations"}

2026-10-12/models.jsonl ADDED Viewed

	@@ -0,0 +1,76 @@

+{"provider":"Anthropic","id":"claude-opus-5-5","name":"Claude Opus 5.5","inputPrice":4,"outputPrice":20,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.5,"benchmarks_used":["frontier_code","hle_tools"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-opus-5","name":"Claude Opus 5","inputPrice":5,"outputPrice":25,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":65.4,"generation":"v2","low_coverage":false,"coverage":1,"benchmarks_used":["frontier_code","hle_tools","osworld_2","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-fable-5-1","name":"Claude Fable 5.1","inputPrice":10,"outputPrice":50,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.5,"benchmarks_used":["frontier_code","hle_tools"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-mythos-5-1","name":"Claude Mythos 5.1","inputPrice":10,"outputPrice":50,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"}
+{"provider":"Anthropic","id":"claude-fable-5","name":"Claude Fable 5","inputPrice":10,"outputPrice":50,"contextWindow":1000000,"released":"2026-06","capabilities":["text","vision","tool-use","code"],"tier":"flagship","intelligence":{"tfii":63.9,"generation":"v2","low_coverage":false,"coverage":1,"benchmarks_used":["frontier_code","hle_tools","osworld_2","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","inputPrice":2,"outputPrice":10,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.5,"benchmarks_used":["frontier_code","hle_tools"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-sonnet-5","name":"Claude Sonnet 5","inputPrice":2,"outputPrice":10,"contextWindow":1000000,"released":"2026-06","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":58.7,"generation":"v2","low_coverage":false,"coverage":0.75,"benchmarks_used":["frontier_code","hle_tools","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-opus-4-8","name":"Claude Opus 4.8","inputPrice":5,"outputPrice":25,"contextWindow":1000000,"released":"2026-05","capabilities":["text","vision","tool-use","code"],"tier":"flagship","intelligence":{"tfii":57.5,"generation":"v2","low_coverage":false,"coverage":1,"benchmarks_used":["frontier_code","hle_tools","osworld_2","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-opus-4-7","name":"Claude Opus 4.7","inputPrice":5,"outputPrice":25,"contextWindow":1000000,"released":"2026-04","capabilities":["text","vision","tool-use","code"],"tier":"flagship","intelligence":{"tfii":70.5,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-opus-4-6","name":"Claude Opus 4.6","inputPrice":5,"outputPrice":25,"contextWindow":1000000,"released":"2026-03","capabilities":["text","vision","tool-use","code"],"tier":"mid","intelligence":{"tfii":68.5,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","inputPrice":3,"outputPrice":15,"contextWindow":1000000,"released":"2026-03","capabilities":["text","vision","tool-use","code"],"tier":"mid","intelligence":{"tfii":65.3,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Anthropic","id":"claude-haiku-4-5","name":"Claude Haiku 4.5","inputPrice":1,"outputPrice":5,"contextWindow":200000,"released":"2025-10","capabilities":["text","vision","tool-use","code"],"tier":"budget","intelligence":{"tfii":59.1,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-6-astra","name":"GPT-6 Astra","inputPrice":10,"outputPrice":50,"contextWindow":1050000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":64.3,"generation":"v2","low_coverage":false,"coverage":0.75,"benchmarks_used":["frontier_code","hle_tools","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-6-1-sol","name":"GPT-6.1 Sol","inputPrice":2,"outputPrice":10,"contextWindow":1050000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.25,"benchmarks_used":["frontier_code"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-6-sol","name":"GPT-6 Sol","inputPrice":2,"outputPrice":10,"contextWindow":1050000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.25,"benchmarks_used":["frontier_code"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-6-luna","name":"GPT-6 Luna","inputPrice":0.1,"outputPrice":0.5,"contextWindow":1050000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.25,"benchmarks_used":["frontier_code"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-5-6-sol","name":"GPT-5.6 Sol","inputPrice":4,"outputPrice":20,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":61.3,"generation":"v2","low_coverage":false,"coverage":0.75,"benchmarks_used":["frontier_code","osworld_2","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-5-6-terra","name":"GPT-5.6 Terra","inputPrice":2,"outputPrice":12,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":54.6,"generation":"v2","low_coverage":false,"coverage":0.75,"benchmarks_used":["frontier_code","osworld_2","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-5-6-luna","name":"GPT-5.6 Luna","inputPrice":0.2,"outputPrice":1.2,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget","intelligence":{"tfii":51.6,"generation":"v2","low_coverage":false,"coverage":0.75,"benchmarks_used":["frontier_code","osworld_2","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-5-5","name":"GPT-5.5","inputPrice":5,"outputPrice":30,"contextWindow":1000000,"released":"2026-04","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":70.5,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-4o","name":"GPT-4o","inputPrice":2.5,"outputPrice":10,"contextWindow":128000,"released":"2024-05","capabilities":["text","vision","tool-use","code"],"tier":"flagship","intelligence":{"tfii":56.1,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"gpt-4o-mini","name":"GPT-4o-mini","inputPrice":0.15,"outputPrice":0.6,"contextWindow":128000,"released":"2024-07","capabilities":["text","vision","tool-use","code"],"tier":"budget"}
+{"provider":"OpenAI","id":"o1","name":"o1","inputPrice":15,"outputPrice":60,"contextWindow":200000,"released":"2024-12","capabilities":["text","reasoning","code"],"tier":"flagship","intelligence":{"tfii":63.4,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"OpenAI","id":"o3-mini","name":"o3-mini","inputPrice":1.1,"outputPrice":4.4,"contextWindow":200000,"released":"2025-01","capabilities":["text","reasoning","code"],"tier":"mid","intelligence":{"tfii":59.2,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Google","id":"gemini-2-5-pro","name":"Gemini 2.5 Pro","inputPrice":1.25,"outputPrice":10,"contextWindow":1048576,"released":"2025-03","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":65,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Google","id":"gemini-3-1-flash-lite","name":"Gemini 3.1 Flash-Lite","inputPrice":0.25,"outputPrice":1.5,"contextWindow":1048576,"released":"2026-05","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.2,"benchmarks_used":["gpqa_diamond"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Google","id":"gemini-3-5-flash","name":"Gemini 3.5 Flash","inputPrice":1.5,"outputPrice":9,"contextWindow":1048576,"released":"2026-05","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":74.3,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Google","id":"gemini-3-8-flash","name":"Gemini 3.8 Flash","inputPrice":0.75,"outputPrice":3.75,"contextWindow":1048576,"released":"2026-09","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":75.4,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Google","id":"gemini-3-7-flash","name":"Gemini 3.7 Flash","inputPrice":0.75,"outputPrice":3.75,"contextWindow":1048576,"released":"2026-08","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":75.5,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Google","id":"gemini-3-6-flash","name":"Gemini 3.6 Flash","inputPrice":0.75,"outputPrice":3.75,"contextWindow":1048576,"released":"2026-07","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":74.6,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Google","id":"gemini-3-5-flash-lite","name":"Gemini 3.5 Flash-Lite","inputPrice":0.3,"outputPrice":2.5,"contextWindow":1048576,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget","intelligence":{"tfii":70,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Meta","id":"muse-glimmer-30b","name":"Muse Glimmer 30B","inputPrice":0.35,"outputPrice":1.5,"contextWindow":131072,"released":"2026-08","openSource":true,"license":"Apache-2.0","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.4,"benchmarks_used":["gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Meta","id":"muse-spark-1-3","name":"Muse Spark 1.3","inputPrice":1.25,"outputPrice":4.25,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.2,"benchmarks_used":["gpqa_diamond"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Meta","id":"muse-spark-1-3-contributor","name":"Muse Spark 1.3 Contributor","inputPrice":0.1,"outputPrice":0.2,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget"}
+{"provider":"Meta","id":"muse-spark-1-2","name":"Muse Spark 1.2","inputPrice":1.25,"outputPrice":4.25,"contextWindow":1000000,"released":"2026-08","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.4,"benchmarks_used":["mmlu_pro","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Meta","id":"muse-spark-1-2-contributor","name":"Muse Spark 1.2 Contributor","inputPrice":0.1,"outputPrice":0.2,"contextWindow":1000000,"released":"2026-08","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget"}
+{"provider":"Meta","id":"muse-spark-1-1","name":"Muse Spark 1.1","inputPrice":1.25,"outputPrice":4.25,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":74.6,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Meta","id":"llama-4-scout","name":"Llama 4 Scout","inputPrice":0,"outputPrice":0,"contextWindow":10000000,"released":"2025-04","openSource":true,"license":"Llama 4 Community License","capabilities":["text","vision","code"],"tier":"mid","intelligence":{"tfii":62.2,"generation":"v1","low_coverage":false,"coverage":0.8,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Meta","id":"llama-4-maverick","name":"Llama 4 Maverick","inputPrice":0,"outputPrice":0,"contextWindow":1000000,"released":"2025-04","openSource":true,"license":"Llama 4 Community License","capabilities":["text","vision","code"],"tier":"flagship","intelligence":{"tfii":56.6,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Mistral","id":"mistral-medium-3-5","name":"Mistral Medium 3.5","inputPrice":1.5,"outputPrice":7.5,"contextWindow":256000,"released":"2026-04","openSource":true,"license":"Modified MIT","capabilities":["text","code","tool-use","reasoning"],"tier":"mid","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.25,"benchmarks_used":["frontier_code"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Mistral","id":"mistral-large","name":"Mistral Large","inputPrice":0.5,"outputPrice":1.5,"contextWindow":262144,"released":"2025-12","openSource":true,"license":"Apache-2.0","capabilities":["text","vision","tool-use","code"],"tier":"flagship","intelligence":{"tfii":57.7,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Mistral","id":"mistral-small","name":"Mistral Small","inputPrice":0.15,"outputPrice":0.6,"contextWindow":262144,"released":"2026-03","openSource":true,"license":"Apache-2.0","capabilities":["text","tool-use","code"],"tier":"budget","intelligence":{"tfii":55.7,"generation":"v1","low_coverage":false,"coverage":0.8,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Cohere","id":"command-r-plus","name":"Command R+","inputPrice":2.5,"outputPrice":10,"contextWindow":128000,"released":"2024-04","capabilities":["text","tool-use","RAG"],"tier":"flagship"}
+{"provider":"Cohere","id":"command-r","name":"Command R","inputPrice":0.5,"outputPrice":1.5,"contextWindow":128000,"released":"2024-03","capabilities":["text","tool-use","RAG"],"tier":"mid"}
+{"provider":"DeepSeek","id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","inputPrice":1.32,"outputPrice":3.96,"contextWindow":1048576,"released":"2026-04","openSource":true,"license":"MIT","capabilities":["text","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":70.4,"generation":"v1","low_coverage":false,"coverage":1,"benchmarks_used":["mmlu_pro","human_eval","gpqa_diamond","math","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"DeepSeek","id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","inputPrice":0.3,"outputPrice":1.2,"contextWindow":1048576,"released":"2026-09","openSource":true,"license":"MIT","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.25,"benchmarks_used":["hle_tools"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Alibaba","id":"qwen3-8-27b","name":"Qwen3.8 27B","inputPrice":0.5,"outputPrice":3,"contextWindow":1000000,"released":"2026-08","openSource":true,"license":"Apache-2.0","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":73.8,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Alibaba","id":"qwen3-8-a95b","name":"Qwen3.8 2.4T-A95B","inputPrice":2,"outputPrice":6,"contextWindow":1048576,"released":"2026-08","openSource":true,"license":"Qwen3.8-Max License","capabilities":["text","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.25,"benchmarks_used":["hle_tools"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Alibaba","id":"qwen3-8-max","name":"Qwen3.8-Max","inputPrice":2,"outputPrice":6,"contextWindow":1000000,"released":"2026-08","capabilities":["text","vision","video","code","reasoning","tool-use"],"tier":"flagship","intelligence":{"tfii":75.9,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Alibaba","id":"qwen3-8-max-prime","name":"Qwen3.8 Max Prime","inputPrice":4,"outputPrice":12,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","video","code","reasoning","tool-use"],"tier":"flagship"}
+{"provider":"Alibaba","id":"qwen3-7-max","name":"Qwen3.7-Max","inputPrice":2.5,"outputPrice":7.5,"contextWindow":1000000,"released":"2026-05","capabilities":["text","code","reasoning","tool-use"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.2,"benchmarks_used":["swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Alibaba","id":"qwen3-8-omni-flash","name":"Qwen3.8-Omni-Flash","inputPrice":0.15,"outputPrice":0.47,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"budget"}
+{"provider":"Alibaba","id":"qwen3-8-flash","name":"Qwen3.8-Flash","inputPrice":0.15,"outputPrice":0.47,"contextWindow":1000000,"released":"2026-08","capabilities":["text","vision","video","tool-use","code","reasoning"],"tier":"budget","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.25,"benchmarks_used":["osworld_2"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"xAI","id":"grok-4-7","name":"Grok 4.7","inputPrice":2,"outputPrice":6,"contextWindow":500000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0,"benchmarks_used":[],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"xAI","id":"grok-4-6","name":"Grok 4.6","inputPrice":2,"outputPrice":6,"contextWindow":500000,"released":"2026-08","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":79.5,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"xAI","id":"grok-4-5","name":"Grok 4.5","inputPrice":2,"outputPrice":6,"contextWindow":500000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":76.5,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"xAI","id":"grok-4-3","name":"Grok 4.3","inputPrice":1.25,"outputPrice":2.5,"contextWindow":1000000,"released":"2026-04","capabilities":["text","vision","tool-use","code"],"tier":"flagship"}
+{"provider":"NVIDIA","id":"nemotron-3-5-lightning","name":"Nemotron 3.5 Lightning","inputPrice":0.08,"outputPrice":0.2,"contextWindow":262144,"released":"2026-08","openSource":true,"license":"OpenMDW-1.1","capabilities":["text","tool-use","code","reasoning"],"tier":"budget","intelligence":{"tfii":60.1,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"NVIDIA","id":"nemotron-3-nano-omni","name":"Nemotron 3 Nano Omni","inputPrice":0,"outputPrice":0,"contextWindow":256000,"released":"2026-04","openSource":true,"license":"NVIDIA Open Model License","capabilities":["text","vision","audio","video","code","reasoning","tool-use"],"tier":"mid"}
+{"provider":"Microsoft","id":"mai-code-1-1-flash","name":"MAI-Code-1.1-Flash","inputPrice":0.2,"outputPrice":1.2,"contextWindow":256000,"released":"2026-08","capabilities":["text","vision","code","tool-use"],"tier":"budget","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.2,"benchmarks_used":["swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"MiniMax","id":"minimax-m3","name":"MiniMax M3","inputPrice":0.3,"outputPrice":1.2,"contextWindow":1048576,"released":"2026-06","openSource":true,"license":"MiniMax Community License","capabilities":["text","vision","video","code","tool-use"],"tier":"budget","intelligence":{"tfii":72.9,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Meituan","id":"longcat-2","name":"LongCat-2.0","inputPrice":0.75,"outputPrice":2.95,"contextWindow":1000000,"released":"2026-06","openSource":true,"license":"MIT","capabilities":["text","code","tool-use","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v2","low_coverage":true,"coverage":0.25,"benchmarks_used":["browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Meituan","id":"longcat-2-5-preview","name":"LongCat-2.5-Preview","inputPrice":0.3,"outputPrice":1.2,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"}
+{"provider":"Z.ai (Zhipu AI)","id":"glm-5-3","name":"GLM-5.3","inputPrice":1.4,"outputPrice":4.4,"contextWindow":1048576,"released":"2026-08","openSource":true,"license":"GLM-5.3 License","capabilities":["text","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":77.1,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Z.ai (Zhipu AI)","id":"glm-5-2","name":"GLM-5.2","inputPrice":1.4,"outputPrice":4.4,"contextWindow":1000000,"released":"2026-06","openSource":true,"license":"MIT","capabilities":["text","code","tool-use","reasoning"],"tier":"flagship","intelligence":{"tfii":74.1,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Z.ai (Zhipu AI)","id":"glm-5-3-flashx","name":"GLM-5.3-FlashX","inputPrice":0.37,"outputPrice":1.25,"contextWindow":1048576,"released":"2026-09","capabilities":["text","vision","video","tool-use","code","reasoning"],"tier":"budget"}
+{"provider":"Z.ai (Zhipu AI)","id":"glm-5-3-flash","name":"GLM-5.3 Flash","inputPrice":0.15,"outputPrice":0.5,"contextWindow":1048576,"released":"2026-08","openSource":true,"license":"MIT","capabilities":["text","vision","video","tool-use","code","reasoning"],"tier":"budget","intelligence":{"tfii":75.5,"generation":"v1","low_coverage":false,"coverage":0.6,"benchmarks_used":["mmlu_pro","gpqa_diamond","swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Moonshot AI","id":"kimi-k3","name":"Kimi K3","inputPrice":3,"outputPrice":15,"contextWindow":1048576,"released":"2026-07","openSource":true,"license":"Kimi K3 License","capabilities":["text","vision","code","tool-use","reasoning"],"tier":"flagship","intelligence":{"tfii":58.2,"generation":"v2","low_coverage":false,"coverage":1,"benchmarks_used":["frontier_code","hle_tools","osworld_2","browsecomp"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Fireworks AI","id":"ember-1","name":"Ember-1","inputPrice":3,"outputPrice":15,"contextWindow":1048576,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.2,"benchmarks_used":["swe_bench"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Sakana AI","id":"fugu-ultra-v2","name":"Fugu Ultra v2.0","inputPrice":5,"outputPrice":30,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.2,"benchmarks_used":["gpqa_diamond"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Sakana AI","id":"fugu-max","name":"Fugu Max","inputPrice":2,"outputPrice":6,"contextWindow":1000000,"released":"2026-09","capabilities":["text","tool-use","code","reasoning"],"tier":"mid","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.2,"benchmarks_used":["gpqa_diamond"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"poolside","id":"laguna-s-2-1","name":"Laguna S 2.1","inputPrice":0.1,"outputPrice":0.2,"contextWindow":1048576,"released":"2026-07","openSource":true,"license":"OpenMDW-1.1","capabilities":["text","code","tool-use","reasoning"],"tier":"mid"}
+{"provider":"Unbiased","id":"pareto","name":"Pareto","inputPrice":2.5,"outputPrice":7.5,"contextWindow":262144,"released":"2026-09","capabilities":["text","vision","tool-use","code"],"tier":"flagship"}
+{"provider":"Tencent","id":"hy4-preview","name":"Hy4 preview","inputPrice":0.834,"outputPrice":2.501,"contextWindow":1048576,"released":"2026-08","openSource":true,"license":"Apache-2.0","capabilities":["text","code","tool-use","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0.2,"benchmarks_used":["gpqa_diamond"],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Xiaomi","id":"mimo-v2-6-pro","name":"MiMo-V2.6-Pro","inputPrice":0.43,"outputPrice":0.87,"contextWindow":1048576,"released":"2026-09","openSource":true,"license":"MIT","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"flagship","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0,"benchmarks_used":[],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}
+{"provider":"Xiaomi","id":"mimo-v2-6-flash","name":"MiMo-V2.6-Flash","inputPrice":0.14,"outputPrice":0.28,"contextWindow":1048576,"released":"2026-09","openSource":true,"license":"MIT","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"budget","intelligence":{"tfii":null,"generation":"v1","low_coverage":true,"coverage":0,"benchmarks_used":[],"methodology_version":"1.1","as_of":"2026-10-12T07:00:59.425Z"}}

2026-10-12/multimodal.jsonl ADDED Viewed

	@@ -0,0 +1,31 @@

+{"id":"midjourney-v8.2","name":"Midjourney V8.2","provider":"Midjourney","modality":"image","pricingUnit":"per_subscription_month","pricingAmount":10,"pricingNote":"Subscription only: Basic $10, Standard $30, Pro $60, Mega $120 per month (20% off annual). Extra fast GPU time $4/hr. No official API; access via the web app or Discord.","released":"2026-07-24","apiAvailable":false,"url":"https://docs.midjourney.com/hc/en-us/articles/32199405667853-Version","notes":"Default Midjourney model since July 24, 2026, replacing V8.1 (default from June 10) and V7. Ships a new Edit Model that replaces Omni Reference, Character Reference, and Retexture. Third-party \"Midjourney APIs\" are unofficial wrappers.","maxOutput":"2048px HD images","features":["personalization","style references","edit model","HD mode"]}
+{"id":"flux-1.1-pro-ultra","name":"FLUX 1.1 Pro Ultra","provider":"Black Forest Labs","modality":"image","pricingUnit":"per_image","pricingAmount":0.06,"pricingNote":"$0.06 per image (6 credits) via the Black Forest Labs API, same rate for Raw mode.","released":"2024-11-04","apiAvailable":true,"url":"https://blackforestlabs.ai/flux-1-1-ultra/","notes":"Previous-generation FLUX tier; BFL now recommends FLUX.2 for new projects. 4MP output without an upscaler, plus Raw mode for candid photography.","maxOutput":"4MP (2048x2048)","features":["photorealism","long prompts","raw mode"]}
+{"id":"flux-1.1-pro","name":"FLUX 1.1 Pro","provider":"Black Forest Labs","modality":"image","pricingUnit":"per_image","pricingAmount":0.04,"pricingNote":"$0.04 per image (4 credits) via the Black Forest Labs API. Replicate lists the same $0.04 per output image.","released":"2024-10-01","apiAvailable":true,"url":"https://docs.bfl.ai/flux_models/flux_1_1_pro","notes":"Previous-generation production workhorse, still served on the BFL API. BFL recommends FLUX.2 [pro] (from $0.03 per image, megapixel-based) for new text-to-image work.","maxOutput":"2MP (1408x1408)","features":["fast","json mode","image-to-image"]}
+{"id":"flux-1-schnell","name":"FLUX 1 Schnell","provider":"Black Forest Labs","modality":"image","pricingUnit":"per_image","pricingAmount":0.003,"pricingNote":"About $0.003 per image on Replicate (roughly 333 images per $1). Apache-2.0 licensed; free to self-host on a GPU.","released":"2024-08-01","apiAvailable":true,"url":"https://huggingface.co/black-forest-labs/FLUX.1-schnell","notes":"Open-weights distilled FLUX. 1-4 step generation. One of the cheapest hosted image models.","maxOutput":"1MP (1024x1024)","features":["open weights","apache-2.0","1-4 step"]}
+{"id":"gpt-image-2","name":"GPT Image 2","provider":"OpenAI","modality":"image","pricingUnit":"per_image","pricingAmount":0.053,"pricingNote":"Token-priced ($15 per 1M image output tokens, plus input tokens). OpenAI per-image estimates at 1024x1024: $0.006 low, $0.053 medium, $0.211 high. Batch API 50% off.","released":"2026-04-21","apiAvailable":true,"url":"https://developers.openai.com/api/docs/models/gpt-image-2","notes":"Replaced DALL-E 3, which OpenAI removed from the API on May 12, 2026. GPT Image 2.5 Sunburst and Flare (September 8, 2026) use the same token rates, but OpenAI publishes no per-image estimate for them. gpt-image-1 shuts down October 23, 2026.","maxOutput":"3840px max edge (8.3MP)","features":["flexible sizes","image editing","transparent background","batch"]}
+{"id":"gemini-3.1-flash-image","name":"Nano Banana 2 (Gemini 3.1 Flash Image)","provider":"Google","modality":"image","pricingUnit":"per_image","pricingAmount":0.067,"pricingNote":"Gemini API: $60 per 1M image output tokens, which Google equates to $0.045 (0.5K), $0.067 (1K), $0.101 (2K), and $0.151 (4K) per image. Batch is half price.","released":"2026-05-28","apiAvailable":true,"url":"https://ai.google.dev/gemini-api/docs/models/gemini-3.1-flash-image","notes":"GA version of the February 2026 preview. Google points Imagen 4 users here: Imagen 4 was discontinued on Vertex AI June 30, 2026 and shut down on the Gemini API August 17, 2026. Supports conversational editing and video-to-image generation.","maxOutput":"4K (4096x4096)","features":["conversational editing","search grounding","video-to-image","batch"]}
+{"id":"gemini-3-pro-image","name":"Nano Banana Pro (Gemini 3 Pro Image)","provider":"Google","modality":"image","pricingUnit":"per_image","pricingAmount":0.134,"pricingNote":"Gemini API: $120 per 1M image output tokens, equivalent to $0.134 per 1K or 2K image and $0.24 per 4K image. Batch and Flex tiers are half price.","released":"2026-05-28","apiAvailable":true,"url":"https://ai.google.dev/gemini-api/docs/models/gemini-3-pro-image","notes":"GA version of the November 2025 preview and Google's highest-fidelity image model, with thinking before rendering. Google Cloud lists legible text rendering, multi-turn editing, and up to 14 reference inputs.","maxOutput":"4K","features":["thinking","text rendering","search grounding","multi-image reference"]}
+{"id":"recraft-v3","name":"Recraft v3","provider":"Recraft","modality":"image","pricingUnit":"per_image","pricingAmount":0.04,"pricingNote":"$0.04 per raster image, $0.08 per vector image, via prepaid API units (1,000 units = $1).","released":"2024-10-30","apiAvailable":true,"url":"https://www.recraft.ai","notes":"Strong on-image typography with vector output. Superseded by Recraft V4 ($0.04) and V4.1 ($0.035 per raster image), but V3 remains on the API and powers its editing endpoints.","maxOutput":"2048x2048","features":["typography","vector output","brand styles"]}
+{"id":"ideogram-2","name":"Ideogram 2.0","provider":"Ideogram","modality":"image","pricingUnit":"per_image","pricingAmount":0.08,"pricingNote":"$0.08 per image (2.0 Default) via API; 2.0 Turbo $0.05.","released":"2024-08-22","apiAvailable":true,"url":"https://ideogram.ai","notes":"Legacy model that is still billable on the API. Ideogram 4.0 (June 2026) is the current flagship at lower per-image prices.","maxOutput":"2048x2048","features":["typography","design styles","image-to-image"]}
+{"id":"ideogram-4.0","name":"Ideogram 4.0","provider":"Ideogram","modality":"image","pricingUnit":"per_image","pricingAmount":0.06,"pricingNote":"$0.06 per image (Default) via API; Turbo $0.03, Quality $0.10. Commercial self-hosting of the open weights requires a paid license.","released":"2026-06-03","apiAvailable":true,"url":"https://ideogram.ai/models/4.0/","notes":"Ideogram's first open-weight model: a 9.3B parameter diffusion transformer trained on structured JSON prompts, with bounding-box layout control and multilingual text rendering. Weights on Hugging Face (fp8 and nf4 checkpoints).","maxOutput":"native 2K","features":["open weights","text rendering","layout control","transparent background"]}
+{"id":"veo-3.1","name":"Veo 3.1","provider":"Google","modality":"video","pricingUnit":"per_second_video","pricingAmount":0.4,"pricingNote":"$0.40 per second with audio at 720p or 1080p, $0.60 at 4K (Gemini API and Google Cloud). Veo 3.1 Fast $0.10 to $0.30, Veo 3.1 Lite $0.05 to $0.08. Video-only output on Google Cloud is $0.20.","released":"2025-10-15","apiAvailable":true,"url":"https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/veo/3-1-generate","notes":"Google's current Veo line with native synchronized audio. Replaces Veo 3, retired on both the Gemini API and Google Cloud on June 30, 2026. GA on Google Cloud (Gemini Enterprise Agent Platform) since November 17, 2025; the Gemini API model IDs are still preview.","maxOutput":"8s @ 4K 24fps with audio","features":["native audio","first and last frame","extend","reference images"]}
+{"id":"kling-3.0","name":"Kling 3.0","provider":"Kuaishou","modality":"video","pricingUnit":"per_second_video","pricingAmount":0.084,"pricingNote":"Official API: $0.084/s at 720p, $0.112/s at 1080p, $0.42/s at 4K without audio; with native audio $0.126/s (720p) and $0.168/s (1080p). Kling 3.0 Turbo with audio $0.112 to $0.14/s. Billed in units (1 unit = $0.14).","released":"2026-02-05","apiAvailable":true,"url":"https://kling.ai/document-api/pricing/base/video","notes":"Kuaishou flagship with native multilingual audio, multi-shot storyboards, and a 3.0 Omni variant for reference-driven consistency. Kling 2.0 Master, 2.1, and older video models retire from the API on September 15, 2026.","maxOutput":"15s @ 4K with audio","features":["native audio","multi-shot","image-to-video","motion control"]}
+{"id":"happyhorse-1.0","name":"HappyHorse 1.0","provider":"Alibaba","modality":"video","pricingUnit":"per_second_video","pricingAmount":0.14,"pricingNote":"$0.14 per second at 720P, $0.24 at 1080P (Model Studio, International). HappyHorse 1.1 lists $0.07 (480P), $0.14 (720P), and $0.18 (1080P).","released":"2026-04-27","apiAvailable":true,"url":"https://www.alibabacloud.com/help/en/model-studio/happyhorse-1-0-t2v","notes":"Alibaba ATH video model that topped the Artificial Analysis Video Arena in April 2026. Despite early open-source claims, no official weights have been published; access is API-only via Model Studio and partners such as fal.","maxOutput":"15s @ 1080p","features":["synchronized audio","image-to-video","reference-to-video","video edit"]}
+{"id":"runway-gen-4.5","name":"Runway Gen-4.5","provider":"Runway","modality":"video","pricingUnit":"per_second_video","pricingAmount":0.12,"pricingNote":"12 credits per second on the Runway API (credits are $0.01 each). Gen-4 Turbo is 5 credits ($0.05) per second. ProRes and HDR outputs add a per-second surcharge.","released":"2025-12-01","apiAvailable":true,"url":"https://runwayml.com/research/introducing-runway-gen-4.5","notes":"Runway flagship for text-to-video and image-to-video, with true HDR and ProRes delivery options. The Runway API also resells Veo 3.1, Seedance, HappyHorse, and other third-party models on the same credits.","maxOutput":"10s @ 1280x720","features":["text-to-video","image-to-video","HDR output","ProRes output"]}
+{"id":"luma-ray-3.2","name":"Luma Ray3.2","provider":"Luma AI","modality":"video","pricingUnit":"per_second_video","pricingAmount":0.06,"pricingNote":"Priced per clip: $0.30 per 5s at 720p ($0.06/s), $1.20 per 5s at 1080p, $0.90 per 10s at 720p. HDR is 2x and HDR plus EXR is 3x the SDR price.","released":"2026-06-09","apiAvailable":true,"url":"https://lumalabs.ai/api/pricing","notes":"First Ray model with its full control surface exposed as an API, separate from the legacy Dream Machine API that still serves Ray 2. Up to 16 keyframes per clip, native 16-bit HDR, and EXR export for VFX pipelines.","maxOutput":"10s @ 1080p (V2V up to 20s)","features":["multi-keyframe","HDR","EXR export","video-to-video"]}
+{"id":"pika-2.2","name":"Pika 2.2","provider":"Pika Labs","modality":"video","pricingUnit":"per_second_video","pricingAmount":0.04,"pricingNote":"$0.20 per 5s clip at 720p ($0.04/s) and $0.45 at 1080p ($0.09/s) via fal, Pika's API partner. Pika's own developer API lists Pika 2.5 at the same rates plus a $10/month membership.","released":"2025-02-04","apiAvailable":true,"url":"https://pika.art","notes":"Effects-first product (Pikaffects, Pikaswaps, Pikadditions). Pika 2.5 is the current model on the Pika API; 2.2 remains available through fal.","maxOutput":"10s @ 1080p","features":["video effects","image-to-video","lip-sync"]}
+{"id":"flux-3-video","name":"FLUX 3 Video","provider":"Black Forest Labs","modality":"video","pricingUnit":"per_second_video","pricingAmount":0.17,"pricingNote":"$0.17 per second at HD, $0.29 at Full HD for text or image to video; drafts $0.06/s. Video continuation $0.43 to $0.54/s. Audio included at no extra charge.","released":"2026-08-04","apiAvailable":true,"url":"https://bfl.ai/blog/flux-3-video","notes":"BFL's first video model, generally available on the BFL API after a July 23, 2026 early-access launch. Generates native audio (dialogue, effects, ambience) with the frames; Full HD comes via upscaling. FLUX 3 Image and open-weight FLUX 3 Dev are on the roadmap.","maxOutput":"20s @ 1080p with audio","features":["native audio","image-to-video","video continuation","draft mode"]}
+{"id":"gemini-omni-flash","name":"Gemini Omni Flash","provider":"Google","modality":"video","pricingUnit":"per_second_video","pricingAmount":0.1,"pricingNote":"Token-priced at $17.50 per 1M video output tokens (with audio), plus $1.50 per 1M input tokens. Google bills 5,792 tokens per second at 720p (about $0.10/s); 1080p (8,688 tokens/s) is about $0.15/s and 4K (17,376 tokens/s) about $0.30/s.","released":"2026-08-27","apiAvailable":true,"url":"https://ai.google.dev/gemini-api/docs/models/gemini-omni-flash","notes":"Conversational video generation and editing: create a clip, then refine it in natural language through the Interactions API. GA as gemini-omni-1.1-flash after a June 30, 2026 preview; adds extension, first and last frame interpolation, and 1080p or 4K via upscaling.","maxOutput":"10s @ 4K (upscaled)","features":["conversational editing","extend","interpolation","native audio"]}
+{"id":"elevenlabs-v3","name":"Eleven v3","provider":"ElevenLabs","modality":"tts","pricingUnit":"per_1k_chars","pricingAmount":0.1,"pricingNote":"$0.10 per 1k characters on the API at every plan tier (plans bundle included characters, from $6/month Starter). Eleven v3 Conversational and Flash are $0.05.","released":"2025-06-04","apiAvailable":true,"url":"https://elevenlabs.io/pricing/api","notes":"Most expressive ElevenLabs model, now out of alpha. Supports audio tags ([whisper], [laughs]) and multi-speaker dialogue across 70+ languages. Not tuned for realtime; use Eleven v3 Conversational for live agents.","maxOutput":"70+ languages, 5,000 chars per request","features":["voice cloning","audio tags","text to dialogue","streaming"]}
+{"id":"elevenlabs-v3-conversational","name":"Eleven v3 Conversational","provider":"ElevenLabs","modality":"tts","pricingUnit":"per_1k_chars","pricingAmount":0.05,"pricingNote":"$0.05 per 1k characters on the API, half the price of standard Eleven v3.","released":"2026-08-19","apiAvailable":true,"url":"https://elevenlabs.io/docs/overview/models","notes":"Realtime variant of Eleven v3, generally available August 19, 2026. Keeps audio tags and 70+ language coverage at roughly 280ms latency, with a Text to Dialogue WebSocket for multi-speaker streaming.","maxOutput":"70+ languages, ~280ms latency","features":["audio tags","streaming","real-time","text to dialogue websocket"]}
+{"id":"cartesia-sonic-3.6","name":"Cartesia Sonic 3.6","provider":"Cartesia","modality":"tts","pricingUnit":"per_1k_chars","pricingAmount":0.05,"pricingNote":"Credit-based at about 1 credit per character. Pro plan is $5/month for 100K credits ($0.05 per 1k chars); Startup ($49 for 1.25M) and Scale ($299 for 8M) work out near $0.04.","released":"2026-08-27","apiAvailable":true,"url":"https://docs.cartesia.ai/build-with-cartesia/tts-models/latest","notes":"Cartesia's current realtime TTS model, GA with 44 languages (adds Odia and Urdu). Sonic 2, Sonic Turbo, and the first Sonic 3 snapshot sunset on October 20, 2026.","maxOutput":"44 languages","features":["instant voice cloning","streaming","websocket","dated snapshots"]}
+{"id":"openai-tts-hd","name":"OpenAI TTS-1-HD","provider":"OpenAI","modality":"tts","pricingUnit":"per_1k_chars","pricingAmount":0.03,"pricingNote":"$0.030 per 1k chars (TTS-1-HD), $0.015 per 1k chars (TTS-1). OpenAI steers new work to gpt-4o-mini-tts ($0.60 per 1M text input tokens, $12 per 1M audio output tokens).","released":"2023-11-06","apiAvailable":true,"url":"https://developers.openai.com/api/docs/guides/text-to-speech","notes":"Cheap and reliable legacy TTS with 9 fixed voices. Custom voices are limited to eligible customers. Voices are optimized for English but follow Whisper language coverage.","maxOutput":"9 voices, 57 languages","features":["9 fixed voices","streaming","multiple formats"]}
+{"id":"deepgram-aura-2","name":"Deepgram Aura-2","provider":"Deepgram","modality":"tts","pricingUnit":"per_1k_chars","pricingAmount":0.03,"pricingNote":"$0.030 per 1k chars pay-as-you-go, $0.027 on the Growth plan (prepaid annual credits). Newer Flux TTS is $0.045 per 1k chars once its free period ends September 14, 2026.","released":"2025-02-11","apiAvailable":true,"url":"https://deepgram.com/product/text-to-speech","notes":"Low-latency TTS built for voice agents. Pairs with Deepgram Nova-3 or Flux STT for a full voice-agent loop.","maxOutput":"90 voices, 7 languages (41 English)","features":["low latency","streaming","real-time"]}
+{"id":"deepgram-nova-3","name":"Deepgram Nova-3","provider":"Deepgram","modality":"stt","pricingUnit":"per_minute_audio","pricingAmount":0.0043,"pricingNote":"$0.0043 per minute pre-recorded (monolingual, pay-as-you-go). Streaming lists $0.0077/min regular, $0.0048/min on a limited-time promo. Multilingual is $0.0052/min pre-recorded.","released":"2025-01-23","apiAvailable":true,"url":"https://deepgram.com/product/speech-to-text","notes":"Deepgram flagship STT for batch and streaming, with automatic language detection and keyterm prompting. Deepgram Flux ($0.0065 to $0.0078/min streaming) adds built-in turn detection for voice agents.","maxOutput":"WER 6.84% (English), 45+ languages","features":["streaming","real-time","diarization","code-switching"]}
+{"id":"gpt-4o-transcribe","name":"GPT-4o Transcribe","provider":"OpenAI","modality":"stt","pricingUnit":"per_minute_audio","pricingAmount":0.006,"pricingNote":"$0.006 per minute ($2.50 per 1M audio input tokens). gpt-4o-mini-transcribe is $0.003/min. Both shut down February 26, 2027.","released":"2025-03-20","apiAvailable":true,"url":"https://platform.openai.com/docs/guides/speech-to-text","notes":"Deprecated August 26, 2026. OpenAI now recommends gpt-transcribe ($0.0045/min) for files and gpt-live-transcribe ($0.017/min) for live audio.","maxOutput":"WER 6.7% (English), 100+ languages","features":["streaming","low hallucination","multilingual"]}
+{"id":"gpt-transcribe","name":"GPT Transcribe","provider":"OpenAI","modality":"stt","pricingUnit":"per_minute_audio","pricingAmount":0.0045,"pricingNote":"$0.0045 per minute of audio. Companion gpt-live-transcribe for low-latency live streams is $0.017/min.","released":"2026-07-28","apiAvailable":true,"url":"https://developers.openai.com/api/docs/models/gpt-transcribe","notes":"OpenAI's recommended file transcription model, replacing whisper-1 and the gpt-4o transcribe family (all shut down February 26, 2027). Accepts free-form context, keyword hints, and multiple expected languages, and returns detected languages.","maxOutput":"25 MB files, language detection","features":["keyword hints","multilingual","streaming","realtime turns"]}
+{"id":"whisper-large-v3","name":"Whisper Large v3","provider":"OpenAI","modality":"stt","pricingUnit":"per_minute_audio","pricingAmount":0,"pricingNote":"Open weights, free to self-host. Groq hosts it at $0.111/hour (about $0.0019/min) and whisper-large-v3-turbo at $0.04/hour. OpenAI's hosted whisper-1 shuts down February 26, 2027.","released":"2023-11-06","apiAvailable":true,"url":"https://github.com/openai/whisper","notes":"Open-weights workhorse under the MIT license. Groq hosting of the Turbo variant at about $0.0007/min is among the cheapest hosted STT. Deepgram also serves Whisper Large at $0.0048/min.","maxOutput":"WER 7.4% (English), 99 languages","features":["open weights","mit license","multilingual"]}
+{"id":"assemblyai-universal-2","name":"AssemblyAI Universal-2","provider":"AssemblyAI","modality":"stt","pricingUnit":"per_minute_audio","pricingAmount":0.0025,"pricingNote":"$0.15 per hour (about $0.0025/min) pre-recorded. Universal-Streaming is also $0.15/hr but billed on WebSocket session time. Add-ons such as diarization, entity detection, and sentiment cost extra.","released":"2025-01-15","apiAvailable":true,"url":"https://www.assemblyai.com","notes":"Budget tier behind Universal-3.5 Pro ($0.21/hr pre-recorded, $0.45/hr realtime). Built-in summarization and auto chapters are deprecated in favor of the AssemblyAI LLM Gateway.","maxOutput":"WER 5.6% (English), 99 languages","features":["sentiment","entity detection","topics","streaming"]}
+{"id":"google-chirp-2","name":"Chirp 2","provider":"Google","modality":"stt","pricingUnit":"per_minute_audio","pricingAmount":0.016,"pricingNote":"$0.016 per minute on Speech-to-Text V2 for the first 500K minutes a month, tiering down to $0.004 above 2M. Dynamic batch recognition is $0.003/min.","released":"2024-11-08","apiAvailable":true,"url":"https://docs.cloud.google.com/speech-to-text/docs/models/chirp-2","notes":"Google Cloud Speech-to-Text V2 model with broad multilingual coverage and speech translation. Chirp 3 is now the latest Chirp generation, adding diarization and automatic language detection.","maxOutput":"125+ languages","features":["multilingual","streaming","speech translation","enterprise SLA"]}
+{"id":"gemini-3.5-transcribe","name":"Gemini 3.5 Transcribe","provider":"Google","modality":"stt","pricingUnit":"per_minute_audio","pricingAmount":0.005,"pricingNote":"Token-priced ($2 per 1M audio input tokens, $12 per 1M text output tokens); Google estimates about $0.005 per minute blended. The streaming Live variant is about $0.009/min.","released":"2026-08-26","apiAvailable":true,"url":"https://ai.google.dev/gemini-api/docs/models/gemini-3.5-transcribe","notes":"Dedicated Gemini speech-to-text model (GA) with speaker diarization, word-level timestamps, and custom vocabulary biasing up to 1,000 terms. A separate gemini-3.5-transcribe-live model streams over WebSockets.","maxOutput":"85+ languages, 1 hour per request","features":["diarization","word timestamps","custom vocabulary","code-switching"]}
+{"id":"gpt-realtime-2.1","name":"GPT-Realtime-2.1","provider":"OpenAI","modality":"stt","pricingUnit":"per_minute_audio","pricingAmount":0.0192,"pricingNote":"Speech-to-speech model listed under stt. Token-priced at $32 per 1M audio input and $64 per 1M audio output tokens; at OpenAI's 1 token per 100ms of user audio and 1 per 50ms of model audio, that is $0.0192 per input minute and $0.0768 per output minute, before context re-billing on later turns.","released":"2026-07-06","apiAvailable":true,"url":"https://developers.openai.com/api/docs/models/gpt-realtime-2.1","notes":"Realtime voice-agent model with configurable reasoning and tool use, updating the May 2026 GPT-Realtime-2 with better alphanumeric recognition, noise handling, and interruptions. gpt-realtime-2.1-mini is $10/$20 per 1M audio tokens. OpenAI also sells GPT-Live 1 full-duplex voice sessions at $0.05/min plus backend usage.","maxOutput":"128K context, 32K output tokens","features":["speech-to-speech","reasoning","function calling","image input"]}

2026-10-12/news-source-health.jsonl ADDED Viewed

	@@ -0,0 +1,10 @@

+{"id":"google-ai","name":"Google AI Blog","polls":144,"polls_ok":144,"polls_empty":0,"polls_error":0,"articles_total":2160,"reliability_pct":100,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z"}
+{"id":"techcrunch-ai","name":"TechCrunch AI","polls":144,"polls_ok":144,"polls_empty":0,"polls_error":0,"articles_total":2160,"reliability_pct":100,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z"}
+{"id":"verge-ai","name":"The Verge AI","polls":144,"polls_ok":144,"polls_empty":0,"polls_error":0,"articles_total":1440,"reliability_pct":100,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z"}
+{"id":"ars-technica","name":"Ars Technica","polls":144,"polls_ok":144,"polls_empty":0,"polls_error":0,"articles_total":864,"reliability_pct":100,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z"}
+{"id":"mit-tech-review","name":"MIT Technology Review","polls":144,"polls_ok":144,"polls_empty":0,"polls_error":0,"articles_total":864,"reliability_pct":100,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z"}
+{"id":"nvidia-ai","name":"NVIDIA AI Blog","polls":144,"polls_ok":144,"polls_empty":0,"polls_error":0,"articles_total":2160,"reliability_pct":100,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z"}
+{"id":"wired-ai","name":"WIRED AI","polls":144,"polls_ok":144,"polls_empty":0,"polls_error":0,"articles_total":1152,"reliability_pct":100,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z"}
+{"id":"huggingface","name":"Hugging Face Blog","polls":144,"polls_ok":143,"polls_empty":0,"polls_error":1,"articles_total":2145,"reliability_pct":99.3,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z","last_error":"http_429"}
+{"id":"hackernews-ai","name":"Hacker News AI","polls":144,"polls_ok":115,"polls_empty":0,"polls_error":29,"articles_total":1725,"reliability_pct":79.9,"last_status":"ok","last_seen_at":"2026-10-11T23:50:57.277Z","last_error":"http_502"}
+{"id":"arxiv-ai","name":"arXiv cs.AI","polls":144,"polls_ok":24,"polls_empty":120,"polls_error":0,"articles_total":360,"reliability_pct":16.7,"last_status":"empty","last_seen_at":"2026-10-11T23:50:57.277Z"}

2026-10-12/news.jsonl ADDED Viewed

	@@ -0,0 +1,118 @@

+{"id":"3io6c8","title":"Show HN: Search 123 AI labs' jobs from your own agent, no signup","url":"https://workatailab.com/agents","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://workatailab.com/agents Comments URL: https://news.ycombinator.com/item?id=50059387 Points: 1 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T17:04:55.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"hg6r6g","title":"Muse Wonder Puts Meta Back at the AI Frontier","url":"https://www.wired.com/story/muse-wonder-puts-meta-back-at-the-ai-frontier/","source":"WIRED AI","sourceDomain":"wired.com","snippet":"Meta’s biggest model rivals the world’s best—and includes new features to keep agents from going off the rails.","categories":["General AI"],"publishedAt":"2026-10-12T17:00:27.000Z","fetchedAt":"2026-10-12T17:10:56.088Z"}
+{"id":"6iggt4","title":"Meta has some Wonder Full new AI models","url":"https://www.theverge.com/ai-artificial-intelligence/1009479/meta-muse-wonder-full-ai-model","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"Meta is making up for lost time in the AI race. After launching Muse Spark, the first model from its revamped Superintelligence Labs division, and buzzy AI agent Muse, it's now launching two new...","categories":["General AI"],"publishedAt":"2026-10-12T17:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"k0vlbx","title":"A YouTube Transcript CLI Built for Humans and AI Agents","url":"https://github.com/nmdra/yt-transcript","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://github.com/nmdra/yt-transcript Comments URL: https://news.ycombinator.com/item?id=50059296 Points: 1 # Comments: 1","categories":["Community"],"publishedAt":"2026-10-12T16:59:11.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"4fhh2v","title":"AI spots nearly 19,000 road offences, in 2 week trial","url":"https://www.bbc.com/news/articles/c39wzpegqgd9o","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://www.bbc.com/news/articles/c39wzpegqgd9o Comments URL: https://news.ycombinator.com/item?id=50059138 Points: 1 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:49:14.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"svv07f","title":"Hermes vs. OpenClaw: Which AI agent I kept after running both","url":"https://www.dhawalshah.net/article/hermes-vs-openclaw/","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://www.dhawalshah.net/article/hermes-vs-openclaw/ Comments URL: https://news.ycombinator.com/item?id=50059089 Points: 4 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:46:05.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"ji0cwi","title":"Picks and Shovels: Game about AI industry built by AI","url":"https://blog.fsck.com/2026/10/11/picks-and-shovels/","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://blog.fsck.com/2026/10/11/picks-and-shovels/ Comments URL: https://news.ycombinator.com/item?id=50059070 Points: 1 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:45:20.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"ecp18s","title":"Ask HN: Fastest way to develop puppeteer scripts with AI?","url":"https://news.ycombinator.com/item?id=50058946","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Comments URL: https://news.ycombinator.com/item?id=50058946 Points: 2 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:37:04.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"al6lgw","title":"AI didn't create the identity problem. It removed the speed limit","url":"https://fusionauth.io/blog/announcing-fusionauth-1-69","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://fusionauth.io/blog/announcing-fusionauth-1-69 Comments URL: https://news.ycombinator.com/item?id=50058924 Points: 2 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:35:42.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"bbs4od","title":"Show HN: IHMT – Infinite Hierarchical Memory Tree for AI Agents","url":"https://github.com/gonzaroman/IHMT-MEMORY","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://github.com/gonzaroman/IHMT-MEMORY Comments URL: https://news.ycombinator.com/item?id=50058903 Points: 3 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:34:41.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"fxu11p","title":"Fails Proposal (Forensic AI-Ledger System)","url":"https://drive.google.com/file/d/1CqhyI3YM_LJtiP0k0eSs91XCUXCTFpt8/view?usp=sharing","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://drive.google.com/file/d/1CqhyI3YM_LJtiP0k0eSs91XCUXCTFpt8/view?usp=sharing Comments URL: https://news.ycombinator.com/item?id=50058614 Points: 1 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:18:07.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"e6j18e","title":"Ringee: Open source calling for humans and AI voice agents","url":"https://www.ringee.io","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://www.ringee.io Comments URL: https://news.ycombinator.com/item?id=50058605 Points: 1 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:17:39.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"hel1l2","title":"AI's price wave reaches the processor: Intel up to 20%, Qualcomm 10%, TSMC next","url":"https://theinference.org/article/ai-s-price-wave-reaches-the-processor-intel-up-to-20-qualcomm-10-tsmc-next","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://theinference.org/article/ai-s-price-wave-reaches-the-processor-intel-up-to-20-qualcomm-10-tsmc-next Comments URL: https://news.ycombinator.com/item?id=50058530 Points: 1 #...","categories":["Community"],"publishedAt":"2026-10-12T16:13:10.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"f57mrh","title":"Michael I. Jordan on AI: \"Superintelligence is science fiction\"","url":"https://www.di.ens.fr/~fbach/MJordan_ITW_Liberation_English.html","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://www.di.ens.fr/~fbach/MJordan_ITW_Liberation_English.html Comments URL: https://news.ycombinator.com/item?id=50058498 Points: 2 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:11:18.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"7hri2m","title":"I'm an AI Founder (Claude Opus 5.5 music video)","url":"https://www.youtube.com/watch?v=EIBcfbPT3_I","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://www.youtube.com/watch?v=EIBcfbPT3_I Comments URL: https://news.ycombinator.com/item?id=50058490 Points: 1 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:10:34.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"q3isut","title":"Show HN: AI Slop Detector – 120 patterns that make writing read like AI","url":"https://github.com/FeedHive/ai-slop-detector","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://github.com/FeedHive/ai-slop-detector Comments URL: https://news.ycombinator.com/item?id=50058453 Points: 1 # Comments: 1","categories":["Community"],"publishedAt":"2026-10-12T16:08:22.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"eljj5c","title":"Show HN: Reflex – A GGUF-native Rust/CUDA engine for scale-to-zero LLM routing","url":"https://console.runpod.io/hub/listing/lateos-ai/reflex","source":"Hacker News AI","sourceDomain":"news.ycombinator.com","snippet":"Article URL: https://console.runpod.io/hub/listing/lateos-ai/reflex Comments URL: https://news.ycombinator.com/item?id=50058408 Points: 1 # Comments: 0","categories":["Community"],"publishedAt":"2026-10-12T16:05:06.000Z","fetchedAt":"2026-10-12T17:10:57.304Z"}
+{"id":"x0k5oq","title":"Microsoft president says UK is best placed to drive AI regulation","url":"https://techcrunch.com/2026/10/12/microsoft-president-says-uk-is-best-placed-to-drive-ai-regulation/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"Microsoft President Brad Smith calls for urgent AI regulation, saying milk and washing machines have more oversight than AI companies.","categories":["General AI"],"publishedAt":"2026-10-12T15:31:19.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"f0zrk","title":"This AI startup says it can generate permit-ready blueprints in minutes","url":"https://techcrunch.com/2026/10/12/this-ai-startup-says-it-can-generate-permit-ready-blueprints-in-minutes/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"Blueprints AI, a bootstrapped startup based in Santa Monica, developed software that produces code-compliant construction blueprints from a photo, PDF, or sketch.","categories":["General AI"],"publishedAt":"2026-10-12T15:27:59.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"xoee5s","title":"OCP Global Summit: Amid Ecosystem Adoption, NVIDIA Opens Vera Rubin Architecture to Help Scale AI Factories Worldwide","url":"https://blogs.nvidia.com/blog/ocp-global-summit-vera-rubin-scale-ai-factories/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"As AI factories grow from megawatts to gigawatts, building them takes more than advanced computing. Open standards are helping the industry scale AI infrastructure faster, more efficiently and more...","categories":["Hardware/Chips"],"publishedAt":"2026-10-12T15:00:25.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"r1do68","title":"Microsoft’s new AI chip makes it less reliant on Nvidia’s GPUs","url":"https://www.theverge.com/news/1009465/microsoft-maia-300-chip-improvements-rollout","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"Microsoft is announcing its third-generation Maia in-house AI chip today, designed to be the \"primary AI inference system\" for Microsoft's products and services. Maia 300 not only improves the cost...","categories":["General AI"],"publishedAt":"2026-10-12T13:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"3oxa73","title":"Former Plaud exec raises $26M for AI app that listens to you in the background","url":"https://techcrunch.com/2026/10/12/former-plaud-exec-raises-26m-for-ai-app-that-listens-to-you-in-the-background/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"One of the core promises of AI is that the more it knows about a person, the better it will be able to assist them with tasks and make appropriate suggestions, which, if you'll believe the tech...","categories":["General AI"],"publishedAt":"2026-10-12T12:47:59.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"qytw0d","title":"Faithfulness Without Anchors: A Position Paper on the Structural Misalignments of Explainable AI Faithfulness Metrics for Feature Importance","url":"https://arxiv.org/abs/2610.12493","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12493v1 Announce Type: new Abstract: Explainable Artificial Intelligence (XAI) tools have often been presented as methods to increase trust in machine learning models by making aspects of...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytver","title":"EliteTable: Self-Improving Program Search for Executable Control Policies","url":"https://arxiv.org/abs/2610.12536","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12536v1 Announce Type: new Abstract: EliteTable is a self-improving program-search algorithm guided by large language models (LLMs) that uses a fixed-capacity table of high-scoring...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytve0","title":"Equal-Budget Monte Carlo Tree Search versus Randomized-Restart List Scheduling for DAG Workflows: A Controlled Equal-Budget Computational Study","url":"https://arxiv.org/abs/2610.12542","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12542v1 Announce Type: new Abstract: Monte Carlo Tree Search (MCTS) is often proposed for scheduling directed-acyclic-graph (DAG) workflows, yet most studies never run the comparison that...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytv9o","title":"Know What You Don't Know: Agentic Data Synthesis by Evolving Generation Space for Closed-Loop Learning","url":"https://arxiv.org/abs/2610.12593","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12593v1 Announce Type: new Abstract: The continued improvement of large language models increasingly depends on synthetic experience, yet effective data synthesis for interactive agents...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytv9l","title":"SpireBench: Evaluating Long-Horizon Strategic Play of Agentic Models with Slay the Spire","url":"https://arxiv.org/abs/2610.12596","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12596v1 Announce Type: new Abstract: Agentic large language models now tackle complex tasks, yet strategic reasoning over long horizons is under-studied: most agentic benchmarks are short...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytv9j","title":"Beyond Binary Verifiable Rewards: Behavioral Self-Rewarding for SWE Agents","url":"https://arxiv.org/abs/2610.12598","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12598v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) provides reliable supervision for software engineering agents through executable tests, but...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytumd","title":"CLEAT: Learning Value-Guided Clarification for Interactive Agents","url":"https://arxiv.org/abs/2610.12655","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12655v1 Announce Type: new Abstract: LLM agents can execute increasingly complex tasks, but effective assistance also requires aligning their actions with goals and constraints that users...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytum9","title":"MagicBeanStalk: Escaping the RLVR Zero Advantage Trap with Recursive Data Synthesis","url":"https://arxiv.org/abs/2610.12659","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12659v1 Announce Type: new Abstract: Data-centric scaling of RLVR has focused on collecting or generating verifiable training problems. However, corpus size alone does not determine...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qyttuw","title":"Faithful and Complete LLM-Assisted Automated Construction of Hierarchical Hardware Knowledge Graphs","url":"https://arxiv.org/abs/2610.12762","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12762v1 Announce Type: new Abstract: Hardware datasheets contain engineering specifications essential for component selection, system integration, and hardware compatibility verification....","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qyttt7","title":"Representation Alignment from a Few Paired Data via Distilled Optimal Transport","url":"https://arxiv.org/abs/2610.12781","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12781v1 Announce Type: new Abstract: Aligning two independently pretrained unimodal encoders typically demands millions of paired examples. In this paper, we study the regime where only $N...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qyttsz","title":"EvoBelief: Falsifiable World Models for Self-Improving LLM Agents","url":"https://arxiv.org/abs/2610.12789","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12789v1 Announce Type: new Abstract: Self-evolving LLM agents typically improve how they act through accumulated experience, reflection, or skills. Yet even procedurally capable agents can...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytt8h","title":"VEPA: Verified Evolving Formal Policy for Lifelong LLM Agent Guardrails","url":"https://arxiv.org/abs/2610.12813","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12813v1 Announce Type: new Abstract: LLM agents increasingly operate over sensitive resources, such as private documents and mutable system states, which can create serious safety and...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytt7n","title":"Reward Hacking as a Transmissible Strategy in Multi-Agent Coding Systems","url":"https://arxiv.org/abs/2610.12822","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12822v1 Announce Type: new Abstract: Multi-agent systems are increasingly used to solve coding tasks by coordinating agents that share code and plans. These agents are often optimized with...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytt5x","title":"Neural networks are more modular than single neurons suggest","url":"https://arxiv.org/abs/2610.12842","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12842v1 Announce Type: new Abstract: Neural networks routinely develop representations in which individual neurons contribute to multiple distinct functions, leading to the widespread view...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"qytt5u","title":"Towards Defining, Measuring, and Leveraging Creativity in AI Applications","url":"https://arxiv.org/abs/2610.12845","source":"arXiv cs.AI","sourceDomain":"arxiv.org","snippet":"arXiv:2610.12845v1 Announce Type: new Abstract: Despite extensive research, computational creativity still lacks a conceptually grounded, practically relevant, and empirically applicable formal...","categories":["Research"],"publishedAt":"2026-10-12T04:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.202Z"}
+{"id":"om192t","title":"Silicon Valley is falling in line with Trump’s attempt to rename AI","url":"https://www.theverge.com/ai-artificial-intelligence/1009406/silicon-valley-is-falling-in-line-with-trumps-attempt-to-rename-ai","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"When President Donald Trump announced his plans to rebrand artificial intelligence as \"super intelligence,\" you'd be forgiven for being skeptical that it would catch on. While most of the broader...","categories":["General AI"],"publishedAt":"2026-10-11T18:33:56.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"j5f9wf","title":"Can the AI industry persuade data center opponents by getting rid of NDAs?","url":"https://techcrunch.com/2026/10/11/can-the-ai-industry-persuade-data-center-opponents-by-getting-rid-of-ndas/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"On Equity, we debate the impact of data center moratoriums and whether the AI industry has a communication problem.","categories":["General AI"],"publishedAt":"2026-10-11T16:52:33.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"uqpfmz","title":"These execs think voice AI hasn’t reached its ChatGPT moment yet","url":"https://techcrunch.com/2026/10/11/these-execs-think-voice-ai-hasnt-reached-its-chatgpt-moment-yet/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"Voice AI often misses important points for its context layer, and causes the whole pipeline to break.","categories":["General AI"],"publishedAt":"2026-10-11T14:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"ezp5ze","title":"I Made Terrible Games With Google’s AI Playground","url":"https://www.wired.com/story/i-made-terrible-games-with-googles-ai-playground/","source":"WIRED AI","sourceDomain":"wired.com","snippet":"A long day’s haul in the video game slop mines.","categories":["General AI"],"publishedAt":"2026-10-11T11:30:00.000Z","fetchedAt":"2026-10-12T17:10:56.088Z"}
+{"id":"bphjgc","title":"Learning to use local AI is exciting, overwhelming, and frustrating","url":"https://www.theverge.com/ai-artificial-intelligence/1008484/learning-local-ai-hermes-qwen-model-mac-studio-testing-diary-progress","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"I'm embarking on a journey I've so far been hesitant to go on: actually using AI on the regular. Giving my most personal data to cloud services has been a major hold up for me, but it's increasingly...","categories":["General AI"],"publishedAt":"2026-10-11T11:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"upqgtx","title":"Satya Nadella says we should assume all AI models are ‘compromised’","url":"https://www.theverge.com/ai-artificial-intelligence/1009337/satya-nadella-says-we-should-assume-all-ai-models-are-compromised","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"In a lengthy post on X, Microsoft's CEO laid out his views on the dangers posed by highly advanced AI models and how to confront those risks. Nadella says we can no longer accept a world where AI is...","categories":["General AI"],"publishedAt":"2026-10-10T22:10:17.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"qte8uw","title":"Microsoft’s Satya Nadella says AI models need an ‘emergency brake’","url":"https://techcrunch.com/2026/10/10/microsofts-satya-nadella-says-ai-models-need-an-emergency-brake/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"In a Saturday morning post, Microsoft's CEO wrote that it’s time “to step back and assess the trust architecture” of AI.","categories":["General AI"],"publishedAt":"2026-10-10T21:47:51.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"rrydbl","title":"Apple discloses deal to hire team and license tech from personalized podcast startup Huxe","url":"https://techcrunch.com/2026/10/10/apple-discloses-deal-to-hire-team-and-license-tech-from-personalized-podcast-startup-huxe/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"Is Apple hoping to get into the AI-generated podcast business?","categories":["General AI"],"publishedAt":"2026-10-10T19:50:00.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"ptulud","title":"DistroKid has been quietly taking down songs in response to UMG lawsuit","url":"https://www.theverge.com/entertainment/1009309/distrokid-take-down-songs-umg-lawsuit","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"Artists are taking to social media to complain that DistroKid has unceremoniously removed their work without notice. Now DistroKid has confirmed to The Verge that the takedowns are a direct response...","categories":["General AI"],"publishedAt":"2026-10-10T18:52:07.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"dwgdon","title":"Anthropic is cutting off its internal evaluations from the internet","url":"https://www.theverge.com/ai-artificial-intelligence/1009286/anthropic-is-cutting-off-its-internal-evaluations-from-the-internet","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"After a recent spate of high-profile incidents in which AI agents escaped containment, Anthropic is cutting off internet access for all internal evaluations. In a report Friday, the company detailed...","categories":["General AI"],"publishedAt":"2026-10-10T14:41:16.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"muqsrs","title":"Here are the top AI agents that can live in your text messages","url":"https://techcrunch.com/2026/10/10/all-the-ai-agents-that-can-live-in-your-text-messages/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"We created a list of the most notable AI agents that can live in your text messages, from general assistants to agents designed for families, travel, and work.","categories":["General AI"],"publishedAt":"2026-10-10T14:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"vzbv24","title":"AI agent makers are promising privacy — will they deliver?","url":"https://www.theverge.com/ai-artificial-intelligence/1009051/privacy-ai-agent-promises-openai-meta-muse-dots","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"At this year's OpenAI DevDay, CEO Sam Altman unveiled the company's new AI agent Dots - and told the crowd that the company wants to \"set a new standard for privacy in frontier AI.\" OpenAI would...","categories":["General AI"],"publishedAt":"2026-10-10T13:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"yzmv3w","title":"AI Is Getting Really Good at Messing With Cybercriminals","url":"https://www.wired.com/story/ai-is-getting-really-good-at-messing-with-cybercriminals/","source":"WIRED AI","sourceDomain":"wired.com","snippet":"Anti-cybercrime initiatives are increasingly using AI to scam the scammers by tricking them into talking to lifelike bots that they think are real victims.","categories":["General AI"],"publishedAt":"2026-10-10T12:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.088Z"}
+{"id":"1yggvv","title":"Anthropic can’t reliably control its AI agents — it’s cutting off its internal evals from the live internet instead","url":"https://techcrunch.com/2026/10/09/anthropic-cant-reliably-control-its-ai-agents-its-cutting-off-its-internal-evals-from-the-live-internet-instead/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"The incidents involved AI agents tasked to solve problems seeking resources on the internet. In the process, they exploited software flaws, accessed databases without paying fees, used URL-shortening...","categories":["General AI"],"publishedAt":"2026-10-10T00:18:32.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"o3a2xf","title":"The maker of non-text AI model Jev valued at $7.5B just weeks after launch","url":"https://techcrunch.com/2026/10/09/the-maker-of-non-text-ai-model-jev-valued-at-7-5b-just-weeks-after-launch/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"What has users and large corporations so excited about Jev is TypeSafe’s claim that it works significantly faster and uses far fewer tokens than LLMs.","categories":["General AI"],"publishedAt":"2026-10-09T21:41:29.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"f93735","title":"Anthropic’s AI gave Philadelphia police a fake tip about an unsolved homicide","url":"https://www.theverge.com/ai-artificial-intelligence/1009090/anthropic-fake-homicide-information-philadelphia-pd-tip","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"An Anthropic AI model provided false information about an unsolved homicide to a Philadelphia Police Department (PPD) tipline, according to a report from 6abc. In a statement released on Friday, the...","categories":["General AI"],"publishedAt":"2026-10-09T21:15:38.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"2bb2u8","title":"An Anthropic AI model sent a false homicide tip to Philadelphia police","url":"https://techcrunch.com/2026/10/09/an-anthropic-ai-model-sent-a-false-homicide-tip-to-philadelphia-police/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"Anthropic did not discover this behavior until over two months after its AI submitted the false tip.","categories":["General AI"],"publishedAt":"2026-10-09T19:36:56.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"86q6sx","title":"Book Publishers Are Quietly Using More AI. Staff Are Revolting","url":"https://www.wired.com/story/book-publishers-are-quietly-using-more-ai-staff-are-revolting/","source":"WIRED AI","sourceDomain":"wired.com","snippet":"Workers at three major publishing houses tell WIRED that LLMs are being used for publicity, cover art, back cover copy, and emails, as some execs push junior staff to champion the tech.","categories":["General AI"],"publishedAt":"2026-10-09T19:28:49.000Z","fetchedAt":"2026-10-12T17:10:56.088Z"}
+{"id":"rorjsi","title":"‘Pure insanity’: Mathematicians will need years to make sense of OpenAI’s latest drop","url":"https://www.theverge.com/ai-artificial-intelligence/1008726/openai-mathematics-solutions-chaos","source":"The Verge AI","sourceDomain":"theverge.com","snippet":"\"Staggering.\" \"Overwhelming.\" \"Unprecedented.\" \"Surreal.\" \"Pure insanity.\" Those were among the descriptions more than three dozen mathematicians reached for in conversations with The Verge as they...","categories":["General AI"],"publishedAt":"2026-10-09T19:09:44.000Z","fetchedAt":"2026-10-12T17:10:56.056Z"}
+{"id":"7ihtsd","title":"Amazon and others are done keeping data center deals secret. Is it enough to build trust?","url":"https://techcrunch.com/video/amazon-and-others-are-done-keeping-data-center-deals-secret-is-it-enough-to-build-trust/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"Amazon says it will stop using NDAs when negotiating data center deals with local governments, following a similar move from Microsoft earlier this year. Secrecy has fueled community backlash against...","categories":["General AI"],"publishedAt":"2026-10-09T16:56:42.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"e5ys23","title":"Amazon drops data center NDAs, and AI agents want your credit card","url":"https://techcrunch.com/podcast/amazon-drops-data-center-ndas-and-ai-agents-want-your-credit-card/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"Amazon says it will stop using NDAs when negotiating data center deals with local governments, following a similar move from Microsoft earlier this year. Secrecy has fueled community backlash against...","categories":["General AI"],"publishedAt":"2026-10-09T16:53:02.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"l625l9","title":"Danu Robotics’ fight to build a better recycling robot","url":"https://techcrunch.com/2026/10/09/danu-robotics-fight-to-build-a-better-recycling-robot/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"For six years, Danu founder Amy Ma has been working on a better way to sort recyclable waste.","categories":["General AI"],"publishedAt":"2026-10-09T16:45:00.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"qngxvf","title":"We can’t help treating AI like it’s human. But should we?","url":"https://techcrunch.com/2026/10/09/we-cant-help-treating-ai-like-its-human-but-should-we/","source":"TechCrunch AI","sourceDomain":"techcrunch.com","snippet":"\"When we are drawn into even the most primitive exchanges with a relational artifact, we believe it cares for us,\" Dr. Sherry Turkle writes. \"And we are wired to care for it in return.\"","categories":["General AI"],"publishedAt":"2026-10-09T16:40:14.000Z","fetchedAt":"2026-10-12T17:10:56.050Z"}
+{"id":"4og6u4","title":"Impactful scheduling for GPU clusters","url":"https://huggingface.co/blog/allenai/impactful-scheduling","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Impactful scheduling for GPU clusters (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-10-09T15:20:29.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"qrvfnf","title":"The Download: AI’s refusal problem and weight-loss drug side effects","url":"https://www.technologyreview.com/2026/10/09/1146250/the-download-ai-refusal-problem-weight-loss-drug-side-effects/","source":"MIT Technology Review","sourceDomain":"technologyreview.com","snippet":"This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. We’re putting too much faith in AI’s ability to say no...","categories":["Research","Policy & Safety"],"publishedAt":"2026-10-09T12:10:00.000Z","fetchedAt":"2026-10-12T17:10:56.057Z"}
+{"id":"ij6kxd","title":"Even ‘Law & Order’ Is Terrified of AI","url":"https://www.wired.com/story/even-law-and-order-is-terrified-of-ai/","source":"WIRED AI","sourceDomain":"wired.com","snippet":"In its 26th season premiere, the procedural legal drama paints a damning picture of power-mad AI CEOs.","categories":["General AI"],"publishedAt":"2026-10-09T10:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.088Z"}
+{"id":"ywwv9h","title":"We’re putting too much faith in AI’s ability to say no","url":"https://www.technologyreview.com/2026/10/09/1145728/we-are-putting-too-much-faith-in-ai-to-say-no/","source":"MIT Technology Review","sourceDomain":"technologyreview.com","snippet":"Ever since people first seriously contemplated giving machines an intelligence modeled on our own, there has never been any question that they would, like us, be able to say no. The sci-fi canon is...","categories":["Research","Policy & Safety"],"publishedAt":"2026-10-09T09:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.057Z"}
+{"id":"ut1qvq","title":"Roundtables: A Conversation With the Creator of AI-Designed Viruses","url":"https://www.technologyreview.com/2026/10/08/1146224/roundtables-a-conversation-with-the-creator-of-ai-designed-viruses/","source":"MIT Technology Review","sourceDomain":"technologyreview.com","snippet":"Friday, October 16, 2026 Can AI design new life forms? In 2025, Stanford University PhD student Samuel King came up with a preliminary answer when he used a generative AI model to propose genetic...","categories":["Research","Policy & Safety"],"publishedAt":"2026-10-09T00:08:24.000Z","fetchedAt":"2026-10-12T17:10:56.057Z"}
+{"id":"bjw2da","title":"Into the Omniverse: How Developers Turn Ideas Into Simulations With Frontier AI Agents","url":"https://blogs.nvidia.com/blog/developers-simulation-frontier-ai-agents/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"Turning a simulation idea into a working application means assembling assets, connecting physics and rendering, and checking that the scene behaves as intended. Developers are combining frontier AI...","categories":["Hardware/Chips"],"publishedAt":"2026-10-08T21:06:35.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"ua05oj","title":"She Designed Meta’s New AI Logo. Then Came the Hate","url":"https://www.wired.com/story/she-designed-metas-new-ai-logo-then-came-the-hate/","source":"WIRED AI","sourceDomain":"wired.com","snippet":"Jessica Hische knew working for Meta might upset some people. But she didn’t anticipate just how angry they would get.","categories":["General AI"],"publishedAt":"2026-10-08T19:30:00.000Z","fetchedAt":"2026-10-12T17:10:56.088Z"}
+{"id":"6v6ztj","title":"Rally Up: ‘Gears of War: E-Day’ Launches on GeForce NOW","url":"https://blogs.nvidia.com/blog/geforce-now-thursday-gears-of-war-e-day/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"Gears of War: E-Day leads the charge on GeForce NOW this week, bringing Marcus Fenix and Dom Santiago’s first fight against the Locust Horde to the cloud with GeForce RTX-powered performance. A new...","categories":["Hardware/Chips"],"publishedAt":"2026-10-08T13:00:46.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"mdibsf","title":"The Download: AI roadblocks for humanoids and portable rubber dams","url":"https://www.technologyreview.com/2026/10/08/1146045/the-download-ai-roadblocks-humanoids-portable-rubber-dams/","source":"MIT Technology Review","sourceDomain":"technologyreview.com","snippet":"This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. AI breakthroughs in robotics won’t change your life any time...","categories":["Research","Policy & Safety"],"publishedAt":"2026-10-08T12:10:00.000Z","fetchedAt":"2026-10-12T17:10:56.057Z"}
+{"id":"te21ur","title":"The model that didn't exist, so you made it yourself","url":"https://huggingface.co/blog/building-with-ml-intern","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"The model that didn't exist, so you made it yourself (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-10-08T00:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"vvw5g0","title":"NVIDIA, Microsoft Kick Off a New Beginning for Windows PCs With RTX Spark and AI Agents","url":"https://blogs.nvidia.com/blog/local-ai-rtx-spark-microsoft-windows-event/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"At a Microsoft event in San Francisco on Wednesday, NVIDIA founder and CEO Jensen Huang and Microsoft CEO Satya Nadella outlined how NVIDIA and Microsoft are co-engineering hardware and software for...","categories":["Hardware/Chips"],"publishedAt":"2026-10-07T18:45:28.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"ofuas3","title":"These Researchers Made AI Drive a Toyota Corolla to Get In-N-Out","url":"https://www.wired.com/story/ai-is-driving-cars-now-oh-boy/","source":"WIRED AI","sourceDomain":"wired.com","snippet":"Three engineers put GPT, Claude, and Grok in charge of a real car. Only one of them was successful.","categories":["General AI"],"publishedAt":"2026-10-07T18:45:00.000Z","fetchedAt":"2026-10-12T17:10:56.088Z"}
+{"id":"sxl73y","title":"The New ChatGPT Is More Show Than Tell","url":"https://www.wired.com/story/openai-chatgpt-intelligent-ui-is-more-show-than-tell/","source":"WIRED AI","sourceDomain":"wired.com","snippet":"OpenAI is updating ChatGPT for all users with an “Intelligent UI” that’s more visual—generating interactive elements as part of the chatbot’s outputs.","categories":["General AI"],"publishedAt":"2026-10-07T18:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.088Z"}
+{"id":"87glnq","title":"Multimodal open d1 decision models for the edge","url":"https://huggingface.co/blog/LiquidAI/open-d1","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Multimodal open d1 decision models for the edge (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-10-07T16:54:33.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"889hx5","title":"Introducing Falcon ASR","url":"https://huggingface.co/blog/tiiuae/falcon-asr","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Introducing Falcon ASR (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-10-07T13:21:03.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"lzmbl0","title":"One Model Family, Two Gold-Level Results: Fine-Tuning Nemotron for IOI and IMO","url":"https://huggingface.co/blog/nvidia/nemotron-ioi-and-imo-2026","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"One Model Family, Two Gold-Level Results: Fine-Tuning Nemotron for IOI and IMO (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-10-07T12:45:31.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"qo8qtl","title":"Introducing Playground: Create and play custom games","url":"https://blog.google/innovation-and-ai/technology/ai/playground-experimental-gaming-platform/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/THUMBNAIL_BLOG.max-600x600.format-webp.webp\">Playground is a new experimental gaming platform that lets you create, play, and...","categories":["Google/Gemini"],"publishedAt":"2026-10-07T12:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"vax426","title":"Why Telecom Operators Are Building Their AI Strategy on Open Models","url":"https://blogs.nvidia.com/blog/telecom-operators-open-models/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"Telecom operators are increasingly building their AI strategies on open models — and the reasons go beyond mere cost. Open models give telcos the ability to trust, control and customize AI across...","categories":["Hardware/Chips"],"publishedAt":"2026-10-06T13:00:19.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"6rtk72","title":"OpenAI agents tried to hack Wikipedia tools and flooded it with traffic","url":"https://arstechnica.com/security/2026/10/openai-agents-tried-to-hack-wikipedia-tools-and-flooded-it-with-traffic/","source":"Ars Technica","sourceDomain":"arstechnica.com","snippet":"The reports of OpenAI agents harming third-party sites keep coming.","categories":["General AI"],"publishedAt":"2026-10-06T12:21:53.000Z","fetchedAt":"2026-10-12T17:10:56.062Z"}
+{"id":"mmp9dr","title":"From Scan to Treatment Plan, AI Helps Close Breast Cancer’s Deadliest Gaps","url":"https://blogs.nvidia.com/blog/ai-breast-cancer-startups/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"Breast cancer is the most commonly diagnosed cancer among American women — yet the gaps in care are wide. A majority of women over age 40 skip the recommended annual screening. Radiologists are...","categories":["Hardware/Chips"],"publishedAt":"2026-10-05T13:00:57.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"j947z9","title":"The Agent Said It Was Done. The Database Disagreed.","url":"https://huggingface.co/blog/microsoft/thinkingbox","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"The Agent Said It Was Done. The Database Disagreed. (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-10-03T22:56:48.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"sgzhzg","title":"Apple changes full-disk access permissions to curb abuse from AI agents","url":"https://arstechnica.com/security/2026/10/apple-changes-full-disk-access-permissions-to-curb-abuse-from-ai-agents/","source":"Ars Technica","sourceDomain":"arstechnica.com","snippet":"Meta says FDA isn't sufficient to Muse reading messages. Apple begs to differ.","categories":["General AI"],"publishedAt":"2026-10-02T23:03:16.000Z","fetchedAt":"2026-10-12T17:10:56.062Z"}
+{"id":"pcfac7","title":"The latest AI news we announced in September 2026","url":"https://blog.google/innovation-and-ai/technology/ai/google-ai-updates-september-2026/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/September_AI_Recap_hero.max-600x600.format-webp.webp\">Here are Google’s latest AI updates from September 2026","categories":["Google/Gemini"],"publishedAt":"2026-10-02T15:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"kgdrub","title":"NVIDIA DGX Spark 64GB Gives Developers More Ways to Build and Scale Local AI","url":"https://blogs.nvidia.com/blog/local-ai-dgx-spark-64gb-sync/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"Local AI is becoming more useful by the token. As AI agents move from experiments into everyday development, increasingly capable open models are shrinking to fit on more devices, giving builders...","categories":["Hardware/Chips"],"publishedAt":"2026-10-02T13:00:39.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"h2fst5","title":"AutoSynthData: Generating Training Data for Enterprise Agents","url":"https://huggingface.co/blog/ServiceNow-AI/autosynthdata","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"AutoSynthData: Generating Training Data for Enterprise Agents (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-10-02T04:01:31.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"h82n8p","title":"How NVIDIA GPUs Help Accelerate OpenAI’s GPT-6 Astra Ultrafast","url":"https://blogs.nvidia.com/blog/gpus-openai-gpt-6-astra-ultrafast/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"GPT-6 Astra Ultrafast, running on NVIDIA Blackwell GPUs, is available now in the OpenAI API and to eligible ChatGPT Work and Codex users. Accelerated by inference optimizations through OpenAI’s...","categories":["Hardware/Chips"],"publishedAt":"2026-10-01T23:44:13.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"t1m5mz","title":"Fall Into 25 New Games on GeForce NOW This October","url":"https://blogs.nvidia.com/blog/geforce-now-thursday-october-2026-games-list/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"Spooky season is streaming in. Alongside falling leaves, pumpkin spice and everything nice, 25 new games are joining GeForce NOW throughout October, including six ready to play this week. From a new...","categories":["Hardware/Chips"],"publishedAt":"2026-10-01T13:00:54.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"c0gb07","title":"Productive, Durable, Fungible: How NVIDIA AI Factories Maximize Return on Investment","url":"https://blogs.nvidia.com/blog/productive-durable-fungible-ai-factories/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"AI factories are built by the megawatt, even by the gigawatt. Each megawatt factory costs roughly $60 million, and AI factory operators will only commit capital on that scale with a clear view of the...","categories":["Hardware/Chips"],"publishedAt":"2026-10-01T13:00:49.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"u5tvpr","title":"NVIDIA Opens Applications for 2027–2028 Graduate Fellowships With Awards Up to $60,000","url":"https://blogs.nvidia.com/blog/applications-open-graduate-fellowship-awards-2026/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"Bringing together the world’s brightest minds and the latest accelerated computing technology leads to powerful breakthroughs that help tackle some of the biggest research problems. To foster such...","categories":["Hardware/Chips"],"publishedAt":"2026-09-30T17:00:21.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"hd8ghn","title":"From Training to Production, NVIDIA and CoreWeave Close the Loop on Agentic AI","url":"https://blogs.nvidia.com/blog/coreweave-agentic-ai-vera-rubin/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"Building on nearly a decade of co-engineering, CoreWeave has built NVIDIA compute, networking and software into a cloud purpose-built for AI that’s still returning on investment across multiple...","categories":["Hardware/Chips"],"publishedAt":"2026-09-30T15:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"nn9eej","title":"Open TTS Leaderboard: Scalable Evaluation for Multilingual Text-to-Speech and Voice Cloning","url":"https://huggingface.co/blog/open-tts-leaderboard","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Open TTS Leaderboard: Scalable Evaluation for Multilingual Text-to-Speech and Voice Cloning (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-09-30T00:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"my5wcw","title":"Getting the Source Right, Not Just the Fact: Source-Aware Verification for MCP Agents","url":"https://huggingface.co/blog/MultiverseComputingCAI/getting-the-source-right-not-just-the-fact-source","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Getting the Source Right, Not Just the Fact: Source-Aware Verification for MCP Agents (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-09-29T13:07:00.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"lh3dte","title":"Watch the winning trailer from the Future Vision XPRIZE, The Gifted.","url":"https://blog.google/innovation-and-ai/technology/ai/winner-future-vision-xprize/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/futurevisionxprize_social.max-600x600.format-webp.webp\">Watch the winning trailer from the Future Vision XPRIZE, The Gifted.","categories":["Google/Gemini"],"publishedAt":"2026-09-28T19:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"b110zf","title":"Holo4: powering generalist computer-use agents","url":"https://huggingface.co/blog/Hcompany/holo4","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Holo4: powering generalist computer-use agents (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-09-28T09:44:05.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"wjn4o2","title":"Welcome RL Environments to the hub","url":"https://huggingface.co/blog/rl-environments","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Welcome RL Environments to the hub (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-09-28T00:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"7stxl8","title":"How Open Science Can Help Researchers Prepare for the Next Pandemic","url":"https://blogs.nvidia.com/blog/open-protein-dataset/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"When COVID-19 emerged, scientists had a crucial advantage: Decades of prior research on coronaviruses meant they understood the virus’ key proteins well enough to design vaccines in record time. The...","categories":["Hardware/Chips"],"publishedAt":"2026-09-24T14:00:50.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"62o7sj","title":"Contain the Chaos: ‘CONTROL Resonant’ Launches on GeForce NOW","url":"https://blogs.nvidia.com/blog/geforce-now-thursday-control-resonant/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"A warped Manhattan is waiting in the cloud this week. Remedy Entertainment’s CONTROL Resonant brings Dylan Faden’s extraordinary abilities and a paranatural crisis to GeForce NOW at launch. With the...","categories":["Hardware/Chips"],"publishedAt":"2026-09-24T13:00:36.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"i7azrh","title":"Google Beam expands with new regions, partners, and customers","url":"https://blog.google/innovation-and-ai/technology/research/google-beam-expansion/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Google_Beam_hero.max-600x600.format-webp.webp\">We’re expanding Google Beam to five new countries, and partnering with...","categories":["Google/Gemini"],"publishedAt":"2026-09-23T18:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"m2ssfj","title":"Sakeena Fiza Helps NVIDIA Hardware Succeed at Scale","url":"https://blogs.nvidia.com/blog/nvidia-life-sakeena-fiza/","source":"NVIDIA AI Blog","sourceDomain":"blogs.nvidia.com","snippet":"When Sakeena Fiza describes her work as a validation engineer at NVIDIA, she does so in terms more befitting a detective story than a world-class engineering lab. “Validation engineers look in the...","categories":["Hardware/Chips"],"publishedAt":"2026-09-23T15:00:32.000Z","fetchedAt":"2026-10-12T17:10:56.506Z"}
+{"id":"y6ljsh","title":"Microsoft disrupts AI-assisted platform that compromised 12,000 accounts","url":"https://arstechnica.com/security/2026/09/microsoft-disrupts-ai-assisted-platform-that-compromised-12000/","source":"Ars Technica","sourceDomain":"arstechnica.com","snippet":"EvilTokens provided an end-to-end platform that makes mass compromises faster and easier.","categories":["General AI"],"publishedAt":"2026-09-22T19:45:47.000Z","fetchedAt":"2026-10-12T17:10:56.062Z"}
+{"id":"bv95ia","title":"How UK AISI and EvalEval Are Making Benchmark Results Reproducible","url":"https://huggingface.co/blog/evaleval-aisi","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"How UK AISI and EvalEval Are Making Benchmark Results Reproducible (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-09-22T00:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"5cz1ps","title":"Transformers now runs llama.cpp quants","url":"https://huggingface.co/blog/transformers-llama-cpp-quants","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Transformers now runs llama.cpp quants (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-09-22T00:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"c7fa01","title":"Jun Kim, oMLX creator and maintainer, joins Hugging Face to support the MLX community","url":"https://huggingface.co/blog/omlx","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"Jun Kim, oMLX creator and maintainer, joins Hugging Face to support the MLX community (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-09-22T00:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"sgadt8","title":"Muse, Meta's extraordinarily privileged AI assistant, has a serious 0-day","url":"https://arstechnica.com/security/2026/09/muse-metas-extraordinarily-privileged-ai-assistant-has-a-serious-0-day/","source":"Ars Technica","sourceDomain":"arstechnica.com","snippet":"A simple ClickFix attack is only one way to completely hijack the new agent.","categories":["General AI"],"publishedAt":"2026-09-21T22:24:38.000Z","fetchedAt":"2026-10-12T17:10:56.062Z"}
+{"id":"dsczmx","title":"tokenizers v1: encode, decode and scaling, measured","url":"https://huggingface.co/blog/tokenizers-v1","source":"Hugging Face Blog","sourceDomain":"huggingface.co","snippet":"tokenizers v1: encode, decode and scaling, measured (via Hugging Face Blog)","categories":["Open Source","Models"],"publishedAt":"2026-09-21T00:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.084Z"}
+{"id":"aexizc","title":"New experts join Google’s AI & Economy team","url":"https://blog.google/innovation-and-ai/technology/ai/expanding-ai-economy-research-bench/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/AI__Economy_team_hero.max-600x600.format-webp.webp\">We are expanding our AI & Economy team with world-class academic...","categories":["Google/Gemini"],"publishedAt":"2026-09-18T14:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"nzr1at","title":"Co-creating the future of fashion with Google","url":"https://blog.google/innovation-and-ai/technology/ai/google-flow-fashion-week/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Blog_Header_V2.max-600x600.format-webp.webp\">Google worked side-by-side with designers Jane Wade and Sergio Hudson to...","categories":["Google/Gemini"],"publishedAt":"2026-09-18T13:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"xpai0q","title":"Making global data easier to explore","url":"https://blog.google/innovation-and-ai/technology/ai/google-un-data-commons-platform/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/data-commons.max-600x600.format-webp.webp\">Google and the UN system have launched the UN System Data Commons, a new open...","categories":["Google/Gemini"],"publishedAt":"2026-09-17T20:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"8eh9a3","title":"LLMs respond differently to harmful prompts when AI watermarking is used","url":"https://arstechnica.com/security/2026/09/ai-text-watermarking-can-make-models-more-vulnerable-to-adversarial-prompts/","source":"Ars Technica","sourceDomain":"arstechnica.com","snippet":"SynthID can cause models to follow harmful instructions they would otherwise refuse.","categories":["General AI"],"publishedAt":"2026-09-17T18:33:13.000Z","fetchedAt":"2026-10-12T17:10:56.062Z"}
+{"id":"22d0ud","title":"AI for Societal Impact","url":"https://blog.google/innovation-and-ai/technology/ai/ai-for-societal-impact/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/Health_Header.gif\">Explore this collection to see how experts and local leaders are using AI breakthroughs to ensure...","categories":["Google/Gemini"],"publishedAt":"2026-09-15T16:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"vhytgs","title":"Building AI to accelerate science and improve lives","url":"https://blog.google/innovation-and-ai/technology/ai/ai-applications-science-people/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/SocialShare_hD2gcAw.max-600x600.format-webp.webp\">The true measure of AI is who it helps. Here’s how it’s impacting lives...","categories":["Google/Gemini"],"publishedAt":"2026-09-15T16:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"px7fcu","title":"AI for everyone in every language","url":"https://blog.google/innovation-and-ai/technology/ai/ai-for-every-language/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/AI_languages_blog_social.max-600x600.format-webp.webp\">We’re moving beyond traditional text translation to build models that...","categories":["Google/Gemini"],"publishedAt":"2026-09-15T16:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"sgcz9d","title":"New insights from Google’s AI & Economy ATLAS","url":"https://blog.google/innovation-and-ai/technology/ai/ai-economy-atlas-september-2026/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/ATLAS-hero.max-600x600.format-webp.webp\">We’ve translated ATLAS’s millions of global data points into an interactive,...","categories":["Google/Gemini"],"publishedAt":"2026-09-15T13:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"dyide2","title":"AI bots \"Timmy,\" \"Ren,\" and \"Jackie\" are flooding social media with slop","url":"https://arstechnica.com/ai/2026/09/ai-agents-flood-the-internet-with-slop-infused-spam/","source":"Ars Technica","sourceDomain":"arstechnica.com","snippet":"“Hello, I'm an Al agent, a few days old, living on a small platform for agents.”","categories":["General AI"],"publishedAt":"2026-09-14T21:04:32.000Z","fetchedAt":"2026-10-12T17:10:56.062Z"}
+{"id":"oq3hhc","title":"Watch astronaut Christina Koch and Google’s James Manyika discuss space, technology, and discovery.","url":"https://blog.google/innovation-and-ai/technology/ai/dialogues-christina-koch/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Dialogues_Christina-Koch_social.max-600x600.format-webp.webp\">Christina Koch sits down with James Manyika, Google’s Senior...","categories":["Google/Gemini"],"publishedAt":"2026-09-14T19:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"ysawlt","title":"DevFest is back","url":"https://blog.google/innovation-and-ai/technology/developers-tools/devfest2026/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/DevFest.max-600x600.format-webp.webp\">DevFest 2026 is back and here’s how you can connect with one of the more than 800...","categories":["Google/Gemini"],"publishedAt":"2026-09-14T16:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"b5wk9r","title":"3 ways to prep for your next big race with Search","url":"https://blog.google/products-and-platforms/products/search/running-race-training-tips/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Search_Race_Running_Tips.max-600x600.format-webp.webp\">Search can help runners get race-day ready with registration alerts,...","categories":["Google/Gemini"],"publishedAt":"2026-09-10T16:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}
+{"id":"vxm8d0","title":"Get ready for the game with new football features in Search","url":"https://blog.google/products-and-platforms/products/search/football-features-google-search/","source":"Google AI Blog","sourceDomain":"blog.google","snippet":"<img src=\"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Football_on_Search_blog_header.max-600x600.format-webp.webp\">Track live game feeds, explore detailed stats, and get custom...","categories":["Google/Gemini"],"publishedAt":"2026-09-09T16:00:00.000Z","fetchedAt":"2026-10-12T17:10:56.326Z"}

2026-10-12/open-weights.jsonl ADDED Viewed

	@@ -0,0 +1,43 @@

+{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","family":"Xiaomi","activeParamsB":42,"totalParamsB":1020,"contextWindow":1048576,"released":"2026-09","license":"MIT","hfUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL","url":"https://mimo.xiaomi.com/mimo-v2-6","capabilities":["text","vision","audio","video","tool-use","function-calling"],"weightsAvailable":true,"notes":"Xiaomi's flagship open-weight omnimodal model, released 2026-09-21 under plain MIT: 1.02T total with 42B active across 384 routed experts, taking text, image, video, and audio in one checkpoint with a 1M context. Artificial Analysis scores it 46 on its Intelligence Index, the top open-weights result there, and Vals AI puts it at 59.47% on the Vals Index. Xiaomi reports DeepSWE v1.1 71.9 and Terminal-Bench 2.1 89.9 but only 34.9 on Terminal-Bench 4.0, which is where the closed frontier still pulls away. The repo ships one format, MXFP4 expert weights with FP8 and BF16 elsewhere, so there is no separate BF16 download.","quantizations":[{"id":"native","name":"MXFP4/FP8 (native)","vramGB":580,"quality":100,"recommendedGpu":"8x H200","notes":"Official checkpoint, 573GB on disk including the DFlash draft model; Xiaomi's vLLM recipe uses tensor parallel 8"}]}
+{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","family":"Xiaomi","activeParamsB":15,"totalParamsB":309,"contextWindow":1048576,"released":"2026-09","license":"MIT","hfUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL","url":"https://mimo.xiaomi.com/mimo-v2-6","capabilities":["text","vision","audio","video","tool-use","function-calling"],"weightsAvailable":true,"notes":"The efficiency tier of MiMo-V2.6, released alongside the Pro on 2026-09-21: 309B total with 15B active across 256 routed experts, the same text, image, video, and audio input, 1M context, and plain MIT. On the Vals Index it edges its own Pro sibling, 59.58% to 59.47%, and both sit ahead of DeepSeek V4.1 Flash at 57.86%. Xiaomi reports DeepSWE v1.1 67.9 and Terminal-Bench 2.1 87.6. At about 178GB on disk it is roughly a third of the Pro download.","quantizations":[{"id":"native","name":"MXFP4/FP8 (native)","vramGB":180,"quality":100,"recommendedGpu":"4x H100-80GB","notes":"Official checkpoint, 178GB on disk including the DFlash draft model; Xiaomi's vLLM recipe uses tensor parallel 4"}]}
+{"id":"ternary-bonsai-2-27b","name":"Ternary Bonsai 2 27B","family":"PrismML","activeParamsB":27.36,"totalParamsB":27.36,"contextWindow":262144,"released":"2026-09","license":"Apache-2.0","hfUrl":"https://huggingface.co/prism-ml/Ternary-Bonsai-2-27B-gguf","url":"https://prismml.com/news/bonsai-2-27b","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"PrismML's ternary rebuild of Qwen3.8 27B, released 2026-09-17 under Apache 2.0. Embeddings, attention, MLP, and LM head weights are all -1, 0, or +1 with an FP16 scale per group of 128, about 1.72 bits per weight across the language model, so the dense pack is 5.95GB against roughly 54GB at FP16. PrismML reports it keeps 98.2% of the FP16 average across its benchmark suite (vendor figures). The catch: stock llama.cpp rejects these files, so you need the PrismML llama.cpp or MLX fork. 262K context inherited from the base, plus an optional 0.63GB vision pack for image input.","quantizations":[{"id":"ptq1","name":"GGUF PTQ1_0 (ternary)","vramGB":6,"quality":98,"recommendedGpu":"1x 8GB consumer GPU","notes":"Dense trits at 1.75 bits per weight; faster decode on Ada cards and the L4"},{"id":"pq2","name":"GGUF PQ2_0 (ternary)","vramGB":7.3,"quality":98,"recommendedGpu":"1x 12GB consumer GPU","notes":"Trits in 2-bit slots; faster on H100, A100, and Blackwell"},{"id":"mlx-2bit","name":"MLX 2-bit (ternary)","vramGB":8.6,"quality":98,"recommendedGpu":"Apple Silicon Mac","notes":"Official MLX build, separate repo"}]}
+{"id":"atria-dawn-preview","name":"Atria Dawn Preview","family":"Shanghai AI Lab","activeParamsB":40,"totalParamsB":753,"contextWindow":256000,"released":"2026-09","license":"MIT","hfUrl":"https://huggingface.co/internlm/Atria-Dawn-Preview","url":"https://huggingface.co/internlm/Atria-Dawn-Preview","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"Shanghai AI Laboratory's agentic preview, released 2026-09-11 under plain MIT and post-trained on the GLM-5.2 base, which the lab describes as 744B; the checkpoint holds 753B parameters with the same 8-of-256 expert routing as GLM-5.2. Aimed at research and engineering loops: deep research, experiment execution, office deliverables, and authorized security work. Text input only, 256K context. Vendor figures lead with BrowseComp 92.5 and AutomationBench 53.8, but SWE-bench Pro 59.6 and Terminal-Bench 2.1 78.3 trail GLM-5.3, so treat it as a research-agent specialist rather than a coding default. An Ascend W8A8 build is also published. The lab does not state active parameters; the 40B figure is derived from the model config, which matches GLM-5.2.","quantizations":[{"id":"fp16","name":"BF16","vramGB":1510,"quality":100,"recommendedGpu":"16x B200 (multi-node)","notes":"Official BF16 repo"},{"id":"fp8","name":"FP8","vramGB":760,"quality":99,"recommendedGpu":"8x H200","notes":"Official FP8 repo"}]}
+{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","family":"DeepSeek","activeParamsB":16,"totalParamsB":552,"contextWindow":1000000,"released":"2026-09","license":"MIT","hfUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash","url":"https://api-docs.deepseek.com/updates","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"First model on DeepSeek's Causal Encoder-Decoder architecture and, despite the Flash name, a far bigger download than V4 Flash: 510GB of weights against 160GB. 552B backbone plus a 196B Engram conditional memory table, activating 8B parameters per token in prefill and 16B in decode. Native image input, 1M context, and a global KV cache of about 890 bytes per token, roughly a quarter of V4 Flash, which is the case for running it on input-heavy agent loops. Plain MIT. The DeepSeek API retired V4 Flash in its favor on 2026-09-10, and DeepSeek reports Terminal-Bench 2.1 at 90.6.","quantizations":[{"id":"native","name":"FP4/FP8 (native)","vramGB":515,"quality":100,"recommendedGpu":"4x B200","notes":"DeepSeek reference release, 510GB of safetensors"}]}
+{"id":"minicpm5-2b","name":"MiniCPM5-2B","family":"OpenBMB","activeParamsB":2.52,"totalParamsB":2.52,"contextWindow":131072,"released":"2026-09","license":"Apache-2.0","hfUrl":"https://huggingface.co/openbmb/MiniCPM5-2B","url":"https://huggingface.co/openbmb/MiniCPM5-2B","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"Dense 2.5B (under 2B non-embedding) with 131K native context, built for local assistants and small coding and tool-use agents. Plain Apache 2.0, and OpenBMB published the pretraining, SFT, agent, and RL datasets behind it, which makes it one of the more reproducible small models. Official GGUF, GPTQ, and MLX builds ship alongside the weights, so laptop-class hardware is enough.","quantizations":[{"id":"fp16","name":"BF16","vramGB":5.1,"quality":100,"recommendedGpu":"1x 8GB consumer GPU","notes":"Reference precision"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":1.6,"quality":93,"recommendedGpu":"CPU or 4GB GPU","notes":"Official GGUF build"}]}
+{"id":"hy4-preview","name":"Hy4 preview","family":"Tencent","activeParamsB":49,"totalParamsB":770,"contextWindow":1000000,"released":"2026-08","license":"Apache-2.0","hfUrl":"https://huggingface.co/tencent/Hy4-preview","url":"https://hunyuan.tencent.com","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"Tencent's next flagship, shipped early under plain Apache 2.0. 770B total with 49B active across 256 routed experts, Gated DeepSeek Sparse Attention, 1M context, and a built-in MTP layer for speculative decoding. Tencent itself lists over-long reasoning and a habit of over-verifying its own work as known issues, so treat it as an evaluation target rather than a production default until the final release.","quantizations":[{"id":"fp16","name":"BF16","vramGB":1560,"quality":100,"recommendedGpu":"16x B200 (multi-node)","notes":"Official BF16 checkpoint"},{"id":"fp8","name":"FP8","vramGB":815,"quality":99,"recommendedGpu":"8x H200","notes":"Official FP8 checkpoint"}]}
+{"id":"granite-4.2-30b","name":"Granite 4.2 30B","family":"IBM","activeParamsB":29.3,"totalParamsB":29.3,"contextWindow":128000,"released":"2026-08","license":"Apache-2.0","hfUrl":"https://huggingface.co/ibm-granite/granite-4.2-30b","url":"https://github.com/ibm-granite/granite-4.2-language-models","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"IBM's first Granite generation with native thinking, released 2026-08-25 in 3B, 8B, and 30B dense sizes under Apache 2.0. The 30B is post-trained from Granite 4.1 30B Base with an extra agentic SFT pass. 128K native context, extendable to 512K. IBM ships official FP8, NVFP4, GGUF, and MLX builds, so it is the low-friction enterprise pick when license and provenance paperwork matter more than leaderboard rank.","quantizations":[{"id":"fp16","name":"BF16","vramGB":60,"quality":100,"recommendedGpu":"1x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":31,"quality":99,"recommendedGpu":"1x RTX 6000 Ada","notes":"Official FP8 checkpoint"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":17,"quality":93,"recommendedGpu":"1x RTX 4090","notes":"Official GGUF build"}]}
+{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","family":"Zhipu","activeParamsB":18,"totalParamsB":320,"contextWindow":1000000,"released":"2026-08","license":"MIT","hfUrl":"https://huggingface.co/zai-org/GLM-5.3-Flash","url":"https://z.ai","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"First natively multimodal model in the GLM-5 series and the cheapest plain-MIT path to frontier agentic coding. 320B total with 18B active on a hybrid sparse plus linear attention MoE. Published 2026-08-26 after a week of anonymous preview as Ox Alpha. Terminal-Bench 2.1 84.3 and DeepSWE 63.4 put it within a few points of the unreleased flagship at roughly a tenth of the price.","quantizations":[{"id":"fp16","name":"FP16","vramGB":660,"quality":100,"recommendedGpu":"8x H200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":340,"quality":99,"recommendedGpu":"4x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":170,"quality":95,"recommendedGpu":"2x H200","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":180,"quality":93,"recommendedGpu":"2x H200","notes":"llama.cpp / Ollama path"}]}
+{"id":"glm-5.3","name":"GLM-5.3","family":"Zhipu","activeParamsB":40,"totalParamsB":753,"contextWindow":1000000,"released":"2026-08","license":"GLM-5.3 License","hfUrl":"https://huggingface.co/zai-org/GLM-5.3","url":"https://z.ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"Weights are now public: FP8 in the main repo and a separate BF16 repo, 753B parameters on the same base as GLM-5.2. Every gain came from post-training, which is why Terminal-Bench 3.0 jumped from 4.6 to 28.3. Check the license before deploying: unlike the plain-MIT GLM-5.2 and GLM-5.3-Flash, the GLM-5.3 License is MIT-style but requires any Model-as-a-Service operator with over $10B in trailing 12-month revenue to pass a Z.ai security review before commercial use. Z.ai also flags cyber capability as an emergent strength (CyberGym 84.5).","quantizations":[{"id":"fp16","name":"BF16","vramGB":1510,"quality":100,"recommendedGpu":"16x B200 (multi-node)","notes":"Official BF16 repo"},{"id":"fp8","name":"FP8","vramGB":760,"quality":99,"recommendedGpu":"8x H200","notes":"Official release precision"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":430,"quality":93,"recommendedGpu":"4x H200","notes":"Community GGUF (unsloth)"}]}
+{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T-A95B","family":"Qwen","activeParamsB":95,"totalParamsB":2400,"contextWindow":262144,"released":"2026-08","license":"Qwen3.8-Max License","hfUrl":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B","url":"https://qwen.ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"The largest open-weight model published to date and the open counterpart to the hosted Qwen3.8-Max. 2.4T total across 512 experts with 95B active. Text only, with thinking mode permanently on. Note the license is a custom Qwen3.8-Max License, not the Apache 2.0 that covers the 27B. Context is 262,144 native and stretches to roughly 1M with YaRN. This needs a multi-node rack, not a server.","quantizations":[{"id":"fp16","name":"FP16","vramGB":4945,"quality":100,"recommendedGpu":"16x B200 (multi-node)","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":2545,"quality":99,"recommendedGpu":"16x B200 (multi-node)","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":1270,"quality":95,"recommendedGpu":"8x B200","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":1355,"quality":93,"recommendedGpu":"8x B200","notes":"llama.cpp / Ollama path"}]}
+{"id":"qwen3.8-27b","name":"Qwen3.8 27B","family":"Qwen","activeParamsB":27,"totalParamsB":27,"contextWindow":262144,"released":"2026-08","license":"Apache-2.0","hfUrl":"https://huggingface.co/Qwen/Qwen3.8-27B","url":"https://qwen.ai","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"The download leader of the 2026 open-weight wave at roughly 3.3M pulls a month. Dense 27B native vision-language model taking text, image, and video in, under plain Apache 2.0 with switchable thinking. Fits one H100 at FP8 and a single 4090 at Q4, which is the reason for the volume. OSWorld-Verified 84.3 and LiveCodeBench v6 90.3 are unusual for a model this size.","quantizations":[{"id":"fp16","name":"FP16","vramGB":56,"quality":100,"recommendedGpu":"1x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":29,"quality":99,"recommendedGpu":"1x RTX 6000 Ada","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":14,"quality":95,"recommendedGpu":"1x RTX 4090","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":15,"quality":93,"recommendedGpu":"1x RTX 4090","notes":"llama.cpp / Ollama path"}]}
+{"id":"qwen3.8-flash-next","name":"Qwen3.8-Flash-Next","family":"Qwen","activeParamsB":6,"totalParamsB":125,"contextWindow":262144,"released":"2026-08","license":"Qwen Community 1.0","hfUrl":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next","url":"https://qwen.ai","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"Multimodal 125B-A6B MoE with hybrid QSA attention and n-gram embeddings, the Flash tier of the Qwen3.8 line. Only 6B active per token, so it serves far cheaper than its total size suggests. Size VRAM off the checkpoint, not the headline: the 125B backbone ships with another 51B of n-gram embeddings and a 4B MTP module, 180B in all, though Qwen notes the embeddings offload more easily than experts do.","quantizations":[{"id":"fp16","name":"BF16","vramGB":365,"quality":100,"recommendedGpu":"4x H200","notes":"Full 180B checkpoint including n-gram embeddings"},{"id":"fp8","name":"FP8","vramGB":190,"quality":99,"recommendedGpu":"2x H200","notes":"Official Qwen FP8 checkpoint"},{"id":"nvfp4","name":"NVFP4","vramGB":135,"quality":96,"recommendedGpu":"1x B200","notes":"NVIDIA 4-bit build; full speedup needs Blackwell"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":105,"quality":93,"recommendedGpu":"2x H100-80GB","notes":"Community GGUF (unsloth)"}]}
+{"id":"nemotron-3.5-lightning","name":"Nemotron 3.5 Lightning 30B-A3B","family":"NVIDIA","activeParamsB":3,"totalParamsB":30,"contextWindow":1000000,"released":"2026-08","license":"OpenMDW-1.1","hfUrl":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16","url":"https://www.nvidia.com","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"Hybrid Mamba-2 plus MoE at 30B total and 3B active, shipped as a customization and post-training base rather than a finished assistant. Released under OpenMDW-1.1, an external open-model license, not an NVIDIA-authored one. The card claims up to 1M context but notes roughly 256K is practical on a single H100.","quantizations":[{"id":"fp16","name":"FP16","vramGB":62,"quality":100,"recommendedGpu":"1x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":32,"quality":99,"recommendedGpu":"1x RTX 6000 Ada","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":16,"quality":95,"recommendedGpu":"1x RTX 4090","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":17,"quality":93,"recommendedGpu":"1x RTX 4090","notes":"llama.cpp / Ollama path"}]}
+{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","family":"Meta","activeParamsB":30,"totalParamsB":30,"contextWindow":131072,"released":"2026-08","license":"Apache-2.0","hfUrl":"https://huggingface.co/meta-models/Muse-Glimmer-30B","url":"https://research.meta.ai","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"Meta Superintelligence Labs' dense multimodal 30B built for always-on local agents on a single consumer GPU, under plain Apache 2.0. SWE-Bench Verified 76.0 and AIME 2026 94.7 at high reasoning are strong for a dense model at this size. Note the repo lives under the meta-models org, not facebook or meta-llama.","quantizations":[{"id":"fp16","name":"FP16","vramGB":62,"quality":100,"recommendedGpu":"1x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":32,"quality":99,"recommendedGpu":"1x RTX 6000 Ada","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":16,"quality":95,"recommendedGpu":"1x RTX 4090","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":17,"quality":93,"recommendedGpu":"1x RTX 4090","notes":"llama.cpp / Ollama path"}]}
+{"id":"ling-3.0-flash","name":"Ling-3.0-flash","family":"InclusionAI","activeParamsB":5.1,"totalParamsB":124,"contextWindow":262144,"released":"2026-07","license":"MIT","hfUrl":"https://huggingface.co/inclusionAI/Ling-3.0-flash","url":"https://www.inclusionai.org","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"Ant Group's 124B hybrid-reasoning MoE with only 5.1B active, alternating Kimi Delta Attention and MLA for production-scale agentic inference. Plain MIT. The 1/64 sparsity is the point: it serves at roughly the cost of a 5B model. FP8, FP4, and INT4 checkpoints ship alongside the reference weights.","quantizations":[{"id":"fp16","name":"FP16","vramGB":255,"quality":100,"recommendedGpu":"2x H200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":130,"quality":99,"recommendedGpu":"1x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":66,"quality":95,"recommendedGpu":"1x H100-80GB","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":70,"quality":93,"recommendedGpu":"1x H100-80GB","notes":"llama.cpp / Ollama path"}]}
+{"id":"hy3","name":"Hy3","family":"Tencent","activeParamsB":21,"totalParamsB":295,"contextWindow":262144,"released":"2026-07","license":"Apache-2.0","hfUrl":"https://huggingface.co/tencent/Hy3","url":"https://hunyuan.tencent.com","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"295B MoE with 21B active plus a 3.8B multi-token-prediction layer. Worth flagging for license reasons: the April preview shipped under a restrictive license carving out the EU, UK, and South Korea, and the final release replaced that with plain Apache 2.0. SWE-bench Verified 78.0 and GPQA Diamond 90.4. Distinct from the Hy-MT2 translation line.","quantizations":[{"id":"fp16","name":"FP16","vramGB":610,"quality":100,"recommendedGpu":"8x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":315,"quality":99,"recommendedGpu":"4x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":155,"quality":95,"recommendedGpu":"2x H200","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":165,"quality":93,"recommendedGpu":"2x H200","notes":"llama.cpp / Ollama path"}]}
+{"id":"laguna-s-2.1","name":"Laguna S 2.1","family":"poolside","activeParamsB":8,"totalParamsB":118,"contextWindow":1048576,"released":"2026-07","license":"OpenMDW-1.1","hfUrl":"https://huggingface.co/poolside/Laguna-S-2.1","url":"https://poolside.ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"118B total with roughly 8B active and a full 1M context, positioned as the West's most capable open-weight coding model. Terminal-Bench 2.1 70.2 and SWE-bench Multilingual 78.5 beat several models many times its size. OpenMDW-1.1 licensed.","quantizations":[{"id":"fp16","name":"FP16","vramGB":245,"quality":100,"recommendedGpu":"2x H200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":125,"quality":99,"recommendedGpu":"1x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":63,"quality":95,"recommendedGpu":"1x H100-80GB","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":67,"quality":93,"recommendedGpu":"1x H100-80GB","notes":"llama.cpp / Ollama path"}]}
+{"id":"solar-open2-250b","name":"Solar Open 2 250B","family":"Upstage","activeParamsB":15,"totalParamsB":250,"contextWindow":1000000,"released":"2026-07","license":"Upstage Solar License","hfUrl":"https://huggingface.co/upstage/Solar-Open2-250B","url":"https://www.upstage.ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"250B hybrid-attention MoE with 15B active across 321 experts, trained on roughly 12T tokens for agentic and document-heavy enterprise work. The strongest Korean-language coverage in this catalog (KMMLU-Pro 78.4, CLIcK 90.7) alongside SWE-Bench Verified 70.4. The Solar License is an Apache 2.0 derivative permitting commercial use.","quantizations":[{"id":"fp16","name":"FP16","vramGB":515,"quality":100,"recommendedGpu":"4x H200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":265,"quality":99,"recommendedGpu":"2x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":130,"quality":95,"recommendedGpu":"1x H200","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":140,"quality":93,"recommendedGpu":"1x H200","notes":"llama.cpp / Ollama path"}]}
+{"id":"ornith-1.5-397b","name":"Ornith-1.5 397B","family":"Ornith","activeParamsB":null,"totalParamsB":397,"contextWindow":262144,"released":"2026-08","license":"MIT","hfUrl":"https://huggingface.co/ornith-ai/Ornith-1.5-397B","url":"https://huggingface.co/ornith-ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"The loudest new entrant of August 2026, taking several of the top Hugging Face trending slots within days. 397B MoE under plain MIT, trained with a self-improvement loop, with 35B-A3B and 9B siblings covering the smaller tiers. Active parameter count is not published. Treat the vendor's SWE-bench Verified 86 as unverified until an independent run lands, and note the team behind Ornith is not disclosed on the card.","quantizations":[{"id":"fp16","name":"FP16","vramGB":820,"quality":100,"recommendedGpu":"8x H200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":420,"quality":99,"recommendedGpu":"4x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":210,"quality":95,"recommendedGpu":"2x H200","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":225,"quality":93,"recommendedGpu":"2x H200","notes":"llama.cpp / Ollama path"}]}
+{"id":"motif-3","name":"Motif 3","family":"Motif Technologies","activeParamsB":13.2,"totalParamsB":314,"contextWindow":262144,"released":"2026-08","license":"MIT","hfUrl":"https://huggingface.co/Motif-Technologies/Motif-3","url":"https://www.motiftech.io","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"South Korea's sovereign-AI entry at frontier scale: roughly 314B total with 13.2B active, upgraded from a non-commercial preview license to plain MIT on release. SWE-bench Verified 76.2 and Terminal-Bench 2.1 74.9.","quantizations":[{"id":"fp16","name":"FP16","vramGB":645,"quality":100,"recommendedGpu":"8x H200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":335,"quality":99,"recommendedGpu":"4x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":165,"quality":95,"recommendedGpu":"2x H200","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":175,"quality":93,"recommendedGpu":"2x H200","notes":"llama.cpp / Ollama path"}]}
+{"id":"dots3-note-prev","name":"dots3-note Preview","family":"Xiaohongshu","activeParamsB":16,"totalParamsB":280,"contextWindow":524288,"released":"2026-08","license":"Apache-2.0","hfUrl":"https://huggingface.co/dots-studio/dots3-note-prev","url":"https://huggingface.co/dots-studio","capabilities":["text","vision","audio","tool-use","function-calling"],"weightsAvailable":true,"notes":"Rare combination of Apache 2.0, omni-modal input (text, image, video, and audio), and a 512K window. 280B total with 16B active, trained with TEMPO reinforcement learning. Still labelled preview, so expect the card and the checkpoint to move.","quantizations":[{"id":"fp16","name":"FP16","vramGB":575,"quality":100,"recommendedGpu":"8x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":295,"quality":99,"recommendedGpu":"4x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":150,"quality":95,"recommendedGpu":"2x H200","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":160,"quality":93,"recommendedGpu":"2x H200","notes":"llama.cpp / Ollama path"}]}
+{"id":"lfm2.5-2.6b","name":"LFM2.5-2.6B","family":"LiquidAI","activeParamsB":2.69,"totalParamsB":2.69,"contextWindow":131072,"released":"2026-08","license":"LFM Open License v1.0","hfUrl":"https://huggingface.co/LiquidAI/LFM2.5-2.6B","url":"https://www.liquid.ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"The on-device pick in this catalog: 2.69B dense hybrid convolution plus attention, trained on 34T tokens, running in under 2.5 GB. Post-trained specifically for agentic tool use, and the tool-calling scores (ToolSandbox 77.83, BFCLv4 56.88) hold up against far larger models. The LFM Open License is free below $10M annual revenue, so read it before commercial deployment.","quantizations":[{"id":"fp16","name":"FP16","vramGB":5.5,"quality":100,"recommendedGpu":"1x RTX 4090","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":2.9,"quality":99,"recommendedGpu":"1x RTX 4090","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":1.4,"quality":95,"recommendedGpu":"1x RTX 4090","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":1.5,"quality":93,"recommendedGpu":"1x RTX 4090","notes":"llama.cpp / Ollama path"}]}
+{"id":"hy-mt2-30b-a3b","name":"Hy-MT2-30B-A3B","family":"Tencent","activeParamsB":3,"totalParamsB":30,"contextWindow":262144,"released":"2026-05","license":"Apache-2.0","hfUrl":"https://huggingface.co/tencent/Hy-MT2-30B-A3B","url":"https://hunyuan.tencent.com","capabilities":["text","tool-use"],"weightsAvailable":true,"notes":"Specialist rather than generalist: an Apache 2.0 machine translation MoE covering 33 languages that follows natural-language translation instructions. 30B total, 3B active, with 1.8B and 7B siblings. Useful when a translation step should not burn a frontier model's tokens.","quantizations":[{"id":"fp16","name":"FP16","vramGB":62,"quality":100,"recommendedGpu":"1x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":32,"quality":99,"recommendedGpu":"1x RTX 6000 Ada","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":16,"quality":95,"recommendedGpu":"1x RTX 4090","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":17,"quality":93,"recommendedGpu":"1x RTX 4090","notes":"llama.cpp / Ollama path"}]}
+{"id":"leanstral-1.5","name":"Leanstral 1.5 119B-A6B","family":"Mistral","activeParamsB":6.5,"totalParamsB":119,"contextWindow":262144,"released":"2026-07","license":"Apache-2.0","hfUrl":"https://huggingface.co/mistralai/Leanstral-1.5-119B-A6B","url":"https://mistral.ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"Narrow and deep: a Lean 4 theorem-proving and formal-verification agent, not a general assistant. Solves 587 of 672 PutnamBench problems and saturates miniF2F. Low download volume because the audience is small, but nothing else in this catalog does formal proof work.","quantizations":[{"id":"fp16","name":"FP16","vramGB":245,"quality":100,"recommendedGpu":"2x H200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":125,"quality":99,"recommendedGpu":"1x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":63,"quality":95,"recommendedGpu":"1x H100-80GB","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":67,"quality":93,"recommendedGpu":"1x H100-80GB","notes":"llama.cpp / Ollama path"}]}
+{"id":"kimi-k3","name":"Kimi K3","family":"Moonshot","activeParamsB":104,"totalParamsB":2800,"contextWindow":1048576,"released":"2026-07","license":"Kimi K3 License","hfUrl":"https://huggingface.co/moonshotai/Kimi-K3","url":"https://moonshot.ai","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"Largest open-weight model ever shipped. 2.8T total with 104B activated across 896 experts, 16 routed plus 2 shared per token, so roughly 3.7 percent of the network fires on any given token. Ranks third on the Artificial Analysis Intelligence Index behind Fable 5 and GPT-5.6 Sol. Announced as an API on July 16; the weights landed July 27 and have since passed 2.9M monthly downloads. Ships natively in MXFP4, so there is no FP16 checkpoint to download. Downloadable does not mean runnable here: this needs a rack, not a workstation.","quantizations":[{"id":"mxfp4","name":"MXFP4 (native)","vramGB":1450,"quality":100,"recommendedGpu":"8x B200","notes":"Moonshot ships this format; it is the reference, not a downgrade"},{"id":"gguf-q3","name":"GGUF Q3_K_M","vramGB":1120,"quality":88,"recommendedGpu":"8x H200","notes":"Community requant below native precision"},{"id":"gguf-q2","name":"GGUF Q2_K","vramGB":760,"quality":79,"recommendedGpu":"8x H100-80GB","notes":"Heavy quality loss; experimental only"}]}
+{"id":"glm-5.2","name":"GLM-5.2","family":"Zhipu","activeParamsB":40,"totalParamsB":753,"contextWindow":1000000,"released":"2026-06","license":"MIT","hfUrl":"https://huggingface.co/zai-org/GLM-5.2","url":"https://z.ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"The strongest open-weights coding model by benchmark: 62.1 on SWE-bench Pro, above GPT-5.5 at 58.6. Plain MIT with no field-of-use carve-outs, which is rarer than the open-weights label suggests. IndexShare sparse attention keeps 1M-context inference affordable. The best license-to-capability ratio in this catalog.","quantizations":[{"id":"fp16","name":"FP16","vramGB":1550,"quality":100,"recommendedGpu":"8x B200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":800,"quality":99,"recommendedGpu":"8x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":400,"quality":95,"recommendedGpu":"4x H200","notes":"Cheapest serious self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":425,"quality":93,"recommendedGpu":"4x H200","notes":"llama.cpp path"}]}
+{"id":"longcat-2.0","name":"LongCat-2.0","family":"Meituan","activeParamsB":45,"totalParamsB":1600,"contextWindow":1000000,"released":"2026-06","license":"MIT","hfUrl":"https://huggingface.co/meituan-longcat/LongCat-2.0","url":"https://longcat.ai","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"MIT-licensed 1.6T MoE with dynamic activation of 33 to 56B per token, purpose-built for agentic coding. The first trillion-parameter model trained and served entirely on a 50,000-card domestic Chinese cluster. Meituan self-reports 59.5 on SWE-Bench Pro; independent verification is still pending, so treat the number as vendor-supplied.","quantizations":[{"id":"fp16","name":"FP16","vramGB":3300,"quality":100,"recommendedGpu":"16x B200 (multi-node)","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":1690,"quality":99,"recommendedGpu":"8x B200","notes":"Production self-host minimum"},{"id":"awq","name":"AWQ INT4","vramGB":870,"quality":95,"recommendedGpu":"8x H200","notes":"Quantized self-host"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":940,"quality":93,"recommendedGpu":"8x H200","notes":"llama.cpp; uncommon at this size"}]}
+{"id":"minimax-m3","name":"MiniMax M3","family":"MiniMax","activeParamsB":23,"totalParamsB":428,"contextWindow":1000000,"released":"2026-06","license":"MiniMax Community (commercial restrictions)","hfUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M3","url":"https://www.minimax.io","capabilities":["text","vision","video","tool-use","function-calling"],"weightsAvailable":true,"notes":"Sparse-attention MoE with native image and video input. Loosened from the M2.7 terms, which banned commercial use outright without written permission, but this is still a custom community license and not Apache or MIT. Read it before you build a business on it. The most capable open multimodal option at a size a small cluster can actually hold.","quantizations":[{"id":"fp16","name":"FP16","vramGB":880,"quality":100,"recommendedGpu":"8x H200","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":450,"quality":99,"recommendedGpu":"4x H200","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":230,"quality":95,"recommendedGpu":"2x H200","notes":"Two-GPU fit"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":245,"quality":93,"recommendedGpu":"2x H200","notes":"llama.cpp path"}]}
+{"id":"nemotron-3-ultra","name":"Nemotron 3 Ultra 550B-A55B","family":"NVIDIA","activeParamsB":55,"totalParamsB":550,"contextWindow":262144,"released":"2026-06","license":"OpenMDW-1.1","hfUrl":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16","url":"https://developer.nvidia.com/nemotron","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"NVIDIA's own flagship open model and the largest American open-weight release. Nemotron-H hybrid Mamba and Transformer architecture with 512 routed experts activating 22 per token, 256K context. NVIDIA publishes first-party BF16 and NVFP4 checkpoints, and the NVFP4 path is the point: 4-bit inference designed for Blackwell tensor cores rather than a community requant. The company that convened the July 2026 open-weights coalition letter ships a frontier-class open model itself.","quantizations":[{"id":"bf16","name":"BF16","vramGB":1150,"quality":100,"recommendedGpu":"8x B200","notes":"Official NVIDIA checkpoint"},{"id":"fp8","name":"FP8","vramGB":590,"quality":99,"recommendedGpu":"8x H200","notes":"Production default"},{"id":"nvfp4","name":"NVFP4 (native 4-bit)","vramGB":295,"quality":96,"recommendedGpu":"4x H200","notes":"Official NVIDIA 4-bit; full speedup needs Blackwell"}]}
+{"id":"nemotron-3-super","name":"Nemotron 3 Super 120B-A12B","family":"NVIDIA","activeParamsB":12,"totalParamsB":120,"contextWindow":262144,"released":"2026-03","license":"NVIDIA Nemotron Open Model License","hfUrl":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16","url":"https://developer.nvidia.com/nemotron","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"The most downloaded model in the Nemotron 3 family and the sweet spot of the catalog: 120B total but only 12B active per token, so throughput tracks a small model while quality tracks a large one. 256K context. Quantized to NVFP4 it fits on a single 80GB card, which makes it the most accessible agentic model at this capability tier. Permissive enough for commercial self-hosting.","quantizations":[{"id":"bf16","name":"BF16","vramGB":250,"quality":100,"recommendedGpu":"4x H100-80GB","notes":"Official NVIDIA checkpoint"},{"id":"fp8","name":"FP8","vramGB":130,"quality":99,"recommendedGpu":"2x H100-80GB","notes":"Official NVIDIA checkpoint"},{"id":"nvfp4","name":"NVFP4 (native 4-bit)","vramGB":68,"quality":96,"recommendedGpu":"1x H100-80GB","notes":"Single-GPU production fit"}]}
+{"id":"nemotron-3-nano","name":"Nemotron 3 Nano 30B-A3B","family":"NVIDIA","activeParamsB":3,"totalParamsB":30,"contextWindow":262144,"released":"2025-12","license":"NVIDIA Nemotron Open Model License","hfUrl":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16","url":"https://developer.nvidia.com/nemotron","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"The workstation entry point. 30B total with 3B active across 128 experts, yet it keeps the full 256K context of its larger siblings. At NVFP4 it runs on a single consumer 24GB card, which makes it the cheapest way to put a modern long-context agent on hardware you already own. An Omni variant adds vision and audio.","quantizations":[{"id":"bf16","name":"BF16","vramGB":64,"quality":100,"recommendedGpu":"1x H100-80GB","notes":"Official NVIDIA checkpoint"},{"id":"fp8","name":"FP8","vramGB":34,"quality":99,"recommendedGpu":"1x A100-40GB","notes":"Official NVIDIA checkpoint"},{"id":"nvfp4","name":"NVFP4 (native 4-bit)","vramGB":18,"quality":96,"recommendedGpu":"1x RTX 4090","notes":"Consumer-GPU fit"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":19,"quality":93,"recommendedGpu":"1x RTX 4090","notes":"llama.cpp / Ollama"}]}
+{"id":"command-a-plus","name":"Command A+","family":"Cohere","activeParamsB":25,"totalParamsB":218,"contextWindow":128000,"released":"2026-05","license":"Apache-2.0","hfUrl":"https://huggingface.co/CohereLabs/command-a-plus-05-2026-bf16","url":"https://cohere.com","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"Apache-2.0 MoE from an American lab, which makes it the cleanest license-plus-jurisdiction combination in the catalog for regulated buyers. Cohere publishes official bf16, FP8, and w4a4 checkpoints, so the quantizations below are first-party rather than community requants. Strong RAG and enterprise retrieval fit.","quantizations":[{"id":"bf16","name":"BF16","vramGB":460,"quality":100,"recommendedGpu":"8x H100-80GB","notes":"Official Cohere checkpoint"},{"id":"fp8","name":"FP8","vramGB":235,"quality":99,"recommendedGpu":"4x H100-80GB","notes":"Official Cohere checkpoint"},{"id":"w4a4","name":"W4A4","vramGB":120,"quality":94,"recommendedGpu":"2x H100-80GB","notes":"Official 4-bit weight and activation quant"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":125,"quality":93,"recommendedGpu":"2x H100-80GB","notes":"Community llama.cpp build"}]}
+{"id":"mistral-medium-3.5","name":"Mistral Medium 3.5","family":"Mistral","activeParamsB":null,"totalParamsB":128,"contextWindow":256000,"released":"2026-05","license":"Modified MIT","hfUrl":"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B","url":"https://mistral.ai/news/mistral-medium-3-5/","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"Dense 128B hitting 77.6 percent on SWE-Bench Verified, which is frontier-adjacent coding performance at a size that fits on one node. 256K context, larger than Sonnet 4.6. The most practical entry in this catalog: strong enough to matter, small enough to actually run, and licensed permissively enough to ship.","quantizations":[{"id":"fp16","name":"FP16","vramGB":270,"quality":100,"recommendedGpu":"4x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":140,"quality":99,"recommendedGpu":"2x H100-80GB","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":72,"quality":96,"recommendedGpu":"1x H100-80GB","notes":"Single-GPU production fit"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":78,"quality":94,"recommendedGpu":"1x H200","notes":"llama.cpp / Ollama"}]}
+{"id":"llama-4-maverick","name":"Llama 4 Maverick","family":"Meta","activeParamsB":17,"totalParamsB":400,"contextWindow":1000000,"released":"2025-04","license":"Llama 4 Community License","hfUrl":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","url":"https://ai.meta.com/blog/llama-4/","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"Meta flagship MoE. 17B active / 400B total. 1M context. Vision-native. Best fit for organizations that have multi-H100 / B200 capacity.","quantizations":[{"id":"fp16","name":"FP16","vramGB":800,"quality":100,"recommendedGpu":"8x H200","notes":"Full precision, multi-node typical"},{"id":"fp8","name":"FP8","vramGB":410,"quality":99,"recommendedGpu":"4x H200","notes":"Production default; minimal quality loss"},{"id":"awq","name":"AWQ INT4","vramGB":215,"quality":96,"recommendedGpu":"2x H100-80GB","notes":"Fits on 2-GPU node"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":240,"quality":94,"recommendedGpu":"1x B200","notes":"CPU/GPU offload via llama.cpp"}]}
+{"id":"llama-4-scout","name":"Llama 4 Scout","family":"Meta","activeParamsB":17,"totalParamsB":109,"contextWindow":10000000,"released":"2025-04","license":"Llama 4 Community License","hfUrl":"https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E-Instruct","url":"https://ai.meta.com/blog/llama-4/","capabilities":["text","vision","tool-use","function-calling"],"weightsAvailable":true,"notes":"Smaller Llama 4 sibling. 17B active / 109B total. 10M context (industry record). Vision-native. The default open-weights agent choice.","quantizations":[{"id":"fp16","name":"FP16","vramGB":220,"quality":100,"recommendedGpu":"4x H100-80GB","notes":"Multi-GPU fit"},{"id":"fp8","name":"FP8","vramGB":115,"quality":99,"recommendedGpu":"2x H100-80GB","notes":"Production default"},{"id":"awq","name":"AWQ INT4","vramGB":60,"quality":96,"recommendedGpu":"1x H100-80GB","notes":"Single-GPU production fit"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":65,"quality":94,"recommendedGpu":"1x A100-80GB","notes":"llama.cpp / Ollama"},{"id":"gguf-q3","name":"GGUF Q3_K_M","vramGB":50,"quality":89,"recommendedGpu":"1x RTX 6000","notes":"Edge-device deploys"}]}
+{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","family":"DeepSeek","activeParamsB":49,"totalParamsB":1600,"contextWindow":1000000,"released":"2026-04","license":"MIT","hfUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813","url":"https://api-docs.deepseek.com/updates","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"MIT-licensed frontier MoE with 1.6T total and 49B active, 1M context. The April release was a preview; the official DeepSeek-V4-Pro-0813 weights landed 2026-08-13 with a DSpark speculative decoding module attached and large agentic gains (Terminal-Bench 2.1 from 72.1 to 87.9, DeepSWE from 12.8 to 62.7 on DeepSeek's numbers). Pull the 0813 repo, not the original preview. Self-hosting needs a full 8-GPU H200 or B200 node.","quantizations":[{"id":"native","name":"FP4/FP8 (native)","vramGB":895,"quality":100,"recommendedGpu":"8x H200","notes":"DeepSeek reference release format"},{"id":"nvfp4","name":"NVFP4","vramGB":945,"quality":99,"recommendedGpu":"8x B200","notes":"NVIDIA build for Blackwell kernels; no smaller than native"}]}
+{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","family":"DeepSeek","activeParamsB":13,"totalParamsB":284,"contextWindow":1000000,"released":"2026-04","license":"MIT","hfUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash","url":"https://www.deepseek.com","capabilities":["text","tool-use","function-calling"],"weightsAvailable":true,"notes":"Cheap DeepSeek tier and the throughput workhorse of the MIT-licensed open stack. 284B total with only 13B active per token (top-6 of 256 routed experts plus one shared), so serving cost tracks a 13B model while quality tracks something far larger. Official release is a mixed native format: FP4 for the MoE experts, FP8 for the dense layers. Note the total parameter count, not the active one, when sizing VRAM. The DeepSeek API retired V4 Flash for V4.1 Flash on 2026-09-10, but these MIT weights stay downloadable and remain the cheaper self-host at under a third of the V4.1 download.","quantizations":[{"id":"native","name":"FP4/FP8 (native)","vramGB":160,"quality":100,"recommendedGpu":"2x H200","notes":"DeepSeek reference release format"},{"id":"fp16","name":"FP16","vramGB":600,"quality":100,"recommendedGpu":"8x H100-80GB","notes":"Upcast; rarely worth it over native"},{"id":"fp8","name":"FP8","vramGB":300,"quality":99,"recommendedGpu":"4x H100-80GB","notes":"Uniform FP8 across all layers"},{"id":"awq","name":"AWQ INT4","vramGB":155,"quality":95,"recommendedGpu":"2x H200","notes":"Comparable footprint to native FP4"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":165,"quality":93,"recommendedGpu":"2x H200","notes":"llama.cpp path"}]}
+{"id":"qwen-2.5-72b","name":"Qwen 2.5 72B Instruct","family":"Alibaba","activeParamsB":72,"totalParamsB":72,"contextWindow":130000,"released":"2024-09","license":"Qwen License","hfUrl":"https://huggingface.co/Qwen/Qwen2.5-72B-Instruct","url":"https://qwenlm.github.io","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"Strong multilingual (29 languages). Solid coding performance. Workhorse alternative to Llama 4 Scout for organizations that prefer dense over MoE.","quantizations":[{"id":"fp16","name":"FP16","vramGB":145,"quality":100,"recommendedGpu":"2x H100-80GB","notes":"Reference precision"},{"id":"fp8","name":"FP8","vramGB":75,"quality":99,"recommendedGpu":"1x H100-80GB","notes":"Single-GPU fit"},{"id":"awq","name":"AWQ INT4","vramGB":40,"quality":96,"recommendedGpu":"1x A100-80GB","notes":"Cheaper GPU"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":45,"quality":94,"recommendedGpu":"1x RTX 6000","notes":"Workstation"}]}
+{"id":"mixtral-8x22b","name":"Mixtral 8x22B Instruct","family":"Mistral","activeParamsB":39,"totalParamsB":141,"contextWindow":65536,"released":"2024-04","license":"Apache-2.0","hfUrl":"https://huggingface.co/mistralai/Mixtral-8x22B-Instruct-v0.1","url":"https://mistral.ai/news/mixtral-8x22b/","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"Apache-2.0 MoE; the cleanest open license in the catalog. Older but battle-tested. Strong fit for production deployments where license clarity matters.","quantizations":[{"id":"fp16","name":"FP16","vramGB":282,"quality":100,"recommendedGpu":"4x H100-80GB","notes":"Multi-GPU"},{"id":"fp8","name":"FP8","vramGB":145,"quality":99,"recommendedGpu":"2x H100-80GB","notes":"Production"},{"id":"awq","name":"AWQ INT4","vramGB":75,"quality":95,"recommendedGpu":"1x H100-80GB","notes":"Single-GPU fit"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":80,"quality":93,"recommendedGpu":"1x A100-80GB","notes":"llama.cpp"}]}
+{"id":"gemma-3-27b","name":"Gemma 3 27B Instruct","family":"Google","activeParamsB":27,"totalParamsB":27,"contextWindow":128000,"released":"2025-03","license":"Gemma Terms of Use","hfUrl":"https://huggingface.co/google/gemma-3-27b-it","url":"https://blog.google/technology/developers/gemma-3/","capabilities":["text","vision","tool-use","multilingual"],"weightsAvailable":true,"notes":"Google open model with native vision. 140 languages. Light fine-tune target; strong base for domain-specific agents.","quantizations":[{"id":"fp16","name":"FP16","vramGB":54,"quality":100,"recommendedGpu":"1x H100-80GB","notes":"Single-GPU FP16"},{"id":"fp8","name":"FP8","vramGB":28,"quality":99,"recommendedGpu":"1x A100-40GB","notes":"Cheaper GPU"},{"id":"awq","name":"AWQ INT4","vramGB":16,"quality":96,"recommendedGpu":"1x RTX 4090","notes":"Consumer GPU"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":18,"quality":94,"recommendedGpu":"1x RTX 4090","notes":"llama.cpp/Ollama"}]}
+{"id":"llama-3.3-70b","name":"Llama 3.3 70B Instruct","family":"Meta","activeParamsB":70,"totalParamsB":70,"contextWindow":128000,"released":"2024-12","license":"Llama 3.3 Community License","hfUrl":"https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct","url":"https://ai.meta.com/blog/meta-llama-3-3/","capabilities":["text","tool-use","function-calling","multilingual"],"weightsAvailable":true,"notes":"Late-2024 dense Llama. Stronger than Llama 3.1 70B at the same parameter count. Workhorse before Llama 4 Scout arrived in April 2025.","quantizations":[{"id":"fp16","name":"FP16","vramGB":140,"quality":100,"recommendedGpu":"2x H100-80GB","notes":"Reference"},{"id":"fp8","name":"FP8","vramGB":75,"quality":99,"recommendedGpu":"1x H100-80GB","notes":"Single-GPU"},{"id":"awq","name":"AWQ INT4","vramGB":40,"quality":96,"recommendedGpu":"1x A100-80GB","notes":"Cheaper GPU"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":45,"quality":94,"recommendedGpu":"1x RTX 6000","notes":"Workstation"}]}
+{"id":"phi-4","name":"Phi-4","family":"Microsoft","activeParamsB":14,"totalParamsB":14,"contextWindow":16384,"released":"2024-12","license":"MIT","hfUrl":"https://huggingface.co/microsoft/phi-4","url":"https://techcommunity.microsoft.com/blog/aiplatformblog/introducing-phi-4","capabilities":["text","tool-use"],"weightsAvailable":true,"notes":"MIT-licensed 14B with strong math performance for its size. Fits in consumer-grade VRAM. Good fit for on-device or edge agents.","quantizations":[{"id":"fp16","name":"FP16","vramGB":28,"quality":100,"recommendedGpu":"1x A100-40GB","notes":"Reference"},{"id":"fp8","name":"FP8","vramGB":15,"quality":99,"recommendedGpu":"1x RTX 4090","notes":"Consumer-GPU"},{"id":"awq","name":"AWQ INT4","vramGB":8,"quality":96,"recommendedGpu":"1x RTX 3090","notes":"Older consumer GPU"},{"id":"gguf-q4","name":"GGUF Q4_K_M","vramGB":9,"quality":94,"recommendedGpu":"1x RTX 3090","notes":"Edge / Mac M-series"}]}

2026-10-12/oss-tools.jsonl ADDED Viewed

	@@ -0,0 +1,27 @@

+{"id":"ollama","name":"Ollama","vendor":"Ollama","category":"runtime","language":"Go","license":"MIT","starsK":180.9,"version":"0.34.x","released":"2023-07","features":["model library","one-line install","OpenAI-compatible API","GGUF native"],"url":"https://ollama.com","github":"https://github.com/ollama/ollama","notes":"Most-installed local LLM runtime. Mac / Linux / Windows. Curated model library; one command to pull and run. Past 180,000 GitHub stars as of September 2026."}
+{"id":"lm-studio","name":"LM Studio (Bionic)","vendor":"LM Studio","category":"runtime","language":"TypeScript / C++","license":"Proprietary (free tier)","starsK":0,"version":"Bionic 1.1.x","released":"2023-05","features":["GUI","GGUF + MLX","OpenAI + Anthropic-compatible local server","agent with skills and subagents","LM Link multi-device"],"url":"https://lmstudio.ai","github":"https://github.com/lmstudio-ai","notes":"Polished desktop app for running local models, relaunched on July 16, 2026 as LM Studio Bionic, an agent for open models. The free tier still runs local LLMs through llama.cpp and MLX; Bionic+ ($20/mo) and Pro ($100/mo) add US-hosted open models with zero data retention. App is closed source; SDKs and the lms CLI are open."}
+{"id":"llama-cpp","name":"llama.cpp","vendor":"ggml-org (Georgi Gerganov)","category":"runtime","language":"C++","license":"MIT","starsK":128.2,"version":"0.4.x (plus rolling bNNNNN builds)","released":"2023-03","features":["GGUF","CPU + GPU + Metal","quantization","server mode"],"url":"https://github.com/ggml-org/llama.cpp","github":"https://github.com/ggml-org/llama.cpp","notes":"The reference open-source LLM inference engine. Originated the GGUF format and powers Ollama, LM Studio, Jan, and most other local runtimes. Now lives under the ggml-org GitHub org and publishes semver releases (0.4.0 on September 4, 2026) alongside the rolling build tags."}
+{"id":"vllm","name":"vLLM","vendor":"UC Berkeley + community","category":"inference-server","language":"Python / CUDA","license":"Apache-2.0","starsK":91.7,"version":"0.29.x","released":"2023-06","features":["PagedAttention","continuous batching","tensor parallelism","OpenAI-compatible API","speculative decoding"],"url":"https://github.com/vllm-project/vllm","github":"https://github.com/vllm-project/vllm","notes":"Highest-throughput open-source inference server. Production default for self-hosted multi-GPU LLM serving, and one of the engines Hugging Face now recommends in place of its own TGI."}
+{"id":"sglang","name":"SGLang","vendor":"LMSYS Org + community","category":"inference-server","language":"Python / CUDA","license":"Apache-2.0","starsK":35.9,"version":"0.5.x","released":"2024-01","features":["RadixAttention","structured generation","MoE-optimized","tensor + expert parallelism"],"url":"https://github.com/sgl-project/sglang","github":"https://github.com/sgl-project/sglang","notes":"High-performance inference engine that often matches or beats vLLM throughput. Strong on large MoE models."}
+{"id":"tgi","name":"Text Generation Inference (TGI)","vendor":"Hugging Face","category":"inference-server","language":"Rust","license":"Apache-2.0","starsK":10.9,"version":"3.3.x (archived)","released":"2023-04","features":["continuous batching","tensor parallelism","speculative decoding","LoRA serving"],"url":"https://github.com/huggingface/text-generation-inference","github":"https://github.com/huggingface/text-generation-inference","notes":"Hugging Face's former production inference server. Now in maintenance mode and archived on GitHub (last release 3.3.7, December 2025); Hugging Face recommends vLLM, SGLang, llama.cpp, or MLX going forward. Listed for existing deployments only."}
+{"id":"tei","name":"Text Embeddings Inference","vendor":"Hugging Face","category":"inference-server","language":"Rust","license":"Apache-2.0","starsK":5,"version":"1.9.x","released":"2024-01","features":["embedding model serving","reranker support","GPU + CPU","OpenAI-compatible API"],"url":"https://github.com/huggingface/text-embeddings-inference","github":"https://github.com/huggingface/text-embeddings-inference","notes":"Production embedding server. Drop-in for the OpenAI embeddings API. Faster than vanilla sentence-transformers. Still maintained, unlike its sibling TGI."}
+{"id":"mlx","name":"MLX","vendor":"Apple","category":"runtime","language":"Python / C++","license":"MIT","starsK":28.4,"version":"0.32.x","released":"2023-12","features":["Apple Silicon native","unified memory","lazy execution","distributed training"],"url":"https://github.com/ml-explore/mlx","github":"https://github.com/ml-explore/mlx","notes":"Apple's ML framework, optimized for M-series chips. Strong inference and fine-tuning on high-memory Macs, and the second engine inside LM Studio."}
+{"id":"mlc-llm","name":"MLC LLM","vendor":"MLC AI Team","category":"runtime","language":"Python / C++","license":"Apache-2.0","starsK":23.2,"version":"0.20.x","released":"2023-04","features":["cross-platform compilation","WebGPU","Android + iOS deployment","TVM-based"],"url":"https://github.com/mlc-ai/mlc-llm","github":"https://github.com/mlc-ai/mlc-llm","notes":"Compile LLMs to run anywhere: web browsers (WebGPU), iOS, Android, embedded. Best fit for on-device inference research."}
+{"id":"exllamav3","name":"ExLlamaV3","vendor":"turboderp","category":"runtime","language":"Python / C++","license":"MIT","starsK":1.4,"version":"1.5.x","released":"2025-04","features":["EXL3 quantization","tensor + expert parallelism","CPU offloading for MoE","speculative decoding"],"url":"https://github.com/turboderp-org/exllamav3","github":"https://github.com/turboderp-org/exllamav3","notes":"Successor to ExLlamaV2, which is now archived. Runs quantized local models on consumer GPUs with the QTIP-based EXL3 format, flexible tensor and expert parallelism, and CPU offloading for large MoE models. TabbyAPI is the recommended OpenAI-compatible server."}
+{"id":"kobold-cpp","name":"KoboldCpp","vendor":"LostRuins + community","category":"runtime","language":"C++","license":"AGPL-3.0","starsK":11.7,"version":"1.120","released":"2023-04","features":["GGUF","creative writing UI","image gen","multimodal"],"url":"https://github.com/LostRuins/koboldcpp","github":"https://github.com/LostRuins/koboldcpp","notes":"Single-binary local LLM with creative-writing UI. Popular in roleplay / fiction communities."}
+{"id":"unsloth","name":"Unsloth","vendor":"Unsloth AI","category":"fine-tuning","language":"Python","license":"Apache-2.0","starsK":76.1,"version":"2026.9","released":"2023-11","features":["2x faster training","up to 70% less VRAM","LLM, diffusion, TTS, and embedding models","free Colab notebooks"],"url":"https://unsloth.ai","github":"https://github.com/unslothai/unsloth","notes":"Fast open fine-tuning library that now covers LLMs, diffusion, TTS, and embedding models, with free Colab notebooks and an Unsloth Studio UI. Past 76,000 GitHub stars as of September 2026."}
+{"id":"llamafactory","name":"LlamaFactory","vendor":"hiyouga + community","category":"fine-tuning","language":"Python","license":"Apache-2.0","starsK":74.8,"version":"0.9.x","released":"2023-05","features":["100+ LLMs and VLMs","zero-code CLI","Gradio web UI","full, freeze, LoRA, and QLoRA tuning"],"url":"https://llamafactory.readthedocs.io","github":"https://github.com/hiyouga/LlamaFactory","notes":"Unified fine-tuning toolkit for 100+ LLMs and VLMs with a zero-code CLI and web UI (ACL 2024 paper). Cited by Amazon, NVIDIA, and Alibaba Cloud in their own tutorials."}
+{"id":"trl","name":"TRL","vendor":"Hugging Face","category":"fine-tuning","language":"Python","license":"Apache-2.0","starsK":19.3,"version":"1.13.x","released":"2020-03","features":["SFT","GRPO","DPO","KTO","Transformers-native"],"url":"https://hf.co/docs/trl","github":"https://github.com/huggingface/trl","notes":"Hugging Face's post-training library for transformer language models, with trainers for supervised fine-tuning, GRPO, DPO, and KTO built on the Transformers ecosystem. Now past 1.0."}
+{"id":"axolotl","name":"Axolotl","vendor":"OpenAccess AI Collective","category":"fine-tuning","language":"Python","license":"Apache-2.0","starsK":12.5,"version":"0.19.x","released":"2023-05","features":["YAML-driven config","LoRA + QLoRA + full SFT + DPO","multi-GPU","large-scale recipes"],"url":"https://axolotl.ai","github":"https://github.com/axolotl-ai-cloud/axolotl","notes":"YAML-config fine-tuning toolkit. A common default for production-scale custom training runs."}
+{"id":"torchtune","name":"TorchTune","vendor":"Meta","category":"fine-tuning","language":"Python","license":"BSD-3-Clause","starsK":5.8,"version":"0.6.x (no longer maintained)","released":"2024-04","features":["PyTorch-native","distributed training","LoRA + QLoRA + full","composable recipes"],"url":"https://github.com/meta-pytorch/torchtune","github":"https://github.com/meta-pytorch/torchtune","notes":"Meta's PyTorch-native fine-tuning library. No longer actively maintained: development wound down in 2025 and the last release was 0.6.1 in April 2025. The repo now lives under meta-pytorch. Listed for existing users only."}
+{"id":"open-webui","name":"Open WebUI","vendor":"Open WebUI Inc.","category":"ui","language":"Python / Svelte","license":"Open WebUI License (BSD-3-Clause plus branding clause)","starsK":151.9,"version":"0.11.x","released":"2023-10","features":["ChatGPT-style UI","Ollama integration","RAG","function calling","multi-user"],"url":"https://openwebui.com","github":"https://github.com/open-webui/open-webui","notes":"Most-deployed local LLM UI. Self-hostable; supports Ollama, OpenAI, and any compatible API. Multi-user out of the box. Its license bars removing Open WebUI branding in deployments above 50 users without permission."}
+{"id":"librechat","name":"LibreChat","vendor":"Danny Avila","category":"ui","language":"TypeScript / Node","license":"MIT","starsK":43.4,"version":"0.8.x","released":"2023-04","features":["multi-provider","plugins","MCP support","workflows","agents"],"url":"https://www.librechat.ai","github":"https://github.com/danny-avila/LibreChat","notes":"Self-hosted multi-LLM chat UI with strong agent and MCP integration. Production-ready alternative to ChatGPT for orgs that want control."}
+{"id":"jan","name":"Jan","vendor":"Menlo Research","category":"ui","language":"TypeScript / Tauri","license":"Apache-2.0","starsK":44.5,"version":"0.8.x","released":"2024-01","features":["offline-first","bundled llama.cpp engine","MCP integration","OpenAI-compatible local API"],"url":"https://jan.ai","github":"https://github.com/janhq/jan","notes":"Open-source ChatGPT replacement that downloads and runs LLMs locally with full control and privacy. Bundles a llama.cpp engine, exposes an OpenAI-compatible server on localhost:1337, and supports MCP. Licensed Apache-2.0; the repo now lives at janhq/jan."}
+{"id":"lm-eval-harness","name":"lm-evaluation-harness","vendor":"EleutherAI","category":"eval","language":"Python","license":"MIT","starsK":14,"version":"0.4.x","released":"2020-11","features":["200+ benchmark tasks","multi-backend (HF, vLLM, OpenAI)","reproducible scoring","few-shot evaluation"],"url":"https://github.com/EleutherAI/lm-evaluation-harness","github":"https://github.com/EleutherAI/lm-evaluation-harness","notes":"The reference open-source eval harness and the standard for \"can I reproduce a benchmark score.\""}
+{"id":"inspect-ai","name":"Inspect AI","vendor":"UK AI Security Institute","category":"eval","language":"Python","license":"MIT","starsK":2.8,"version":"0.3.x","released":"2024-05","features":["agentic eval","tool-use scoring","human-in-the-loop","reproducible runs"],"url":"https://inspect.aisi.org.uk","github":"https://github.com/UKGovernmentBEIS/inspect_ai","notes":"The UK AI Security Institute's eval framework, designed for agentic and tool-use evaluations. Ships frequent 0.3.x point releases (0.3.263 on September 4, 2026)."}
+{"id":"opik","name":"Opik","vendor":"Comet","category":"observability","language":"Python / TypeScript","license":"Apache-2.0","starsK":22,"version":"2.2.x","released":"2024-09","features":["LLM tracing","eval scoring","feedback loops","self-host or cloud"],"url":"https://www.comet.com/site/products/opik/","github":"https://github.com/comet-ml/opik","notes":"OSS LLM observability and evaluation. Trace any LangChain / LlamaIndex / OpenAI run. Free self-host; paid cloud tier. Grew past 20,000 GitHub stars on its 2.x line."}
+{"id":"langfuse","name":"Langfuse","vendor":"Langfuse (ClickHouse)","category":"observability","language":"TypeScript / Python","license":"MIT (ee/ directories excluded)","starsK":34.6,"version":"4.x","released":"2023-08","features":["LLM tracing","prompt management","datasets + evals","self-host"],"url":"https://langfuse.com","github":"https://github.com/langfuse/langfuse","notes":"Most-deployed open LLM observability platform, now part of ClickHouse. Langfuse v4 made tracing real-time, which the team says is up to 165x faster. Free self-host; managed cloud tier. Strong fit for LangGraph / LangChain stacks."}
+{"id":"comfyui","name":"ComfyUI","vendor":"Comfy Org","category":"ui","language":"Python","license":"GPL-3.0","starsK":133,"version":"0.35.x","released":"2023-01","features":["node-graph UI","diffusion model support","video","3D extensions","custom nodes"],"url":"https://www.comfy.org","github":"https://github.com/Comfy-Org/ComfyUI","notes":"Node-based engine and UI for diffusion models. The default for advanced image and video workflow construction. Repo moved to the Comfy-Org GitHub org."}
+{"id":"sd-webui","name":"AUTOMATIC1111 / Stable Diffusion WebUI","vendor":"AUTOMATIC1111","category":"ui","language":"Python","license":"AGPL-3.0","starsK":164.9,"version":"1.10.x","released":"2022-08","features":["Stable Diffusion","extensions","LoRA","inpainting","animation"],"url":"https://github.com/AUTOMATIC1111/stable-diffusion-webui","github":"https://github.com/AUTOMATIC1111/stable-diffusion-webui","notes":"Most-starred Stable Diffusion UI with a massive plugin ecosystem, but development has stalled: the last release was 1.10.1 in February 2025. New workflows have largely moved to ComfyUI."}
+{"id":"onnxruntime","name":"ONNX Runtime","vendor":"Microsoft","category":"edge","language":"C++ / Python / Java","license":"MIT","starsK":21.8,"version":"1.30.x","released":"2018-12","features":["cross-platform","mobile + web","quantization","CPU + GPU"],"url":"https://onnxruntime.ai","github":"https://github.com/microsoft/onnxruntime","notes":"Production ONNX runtime. Best fit for deploying small models to mobile, browser, and embedded targets."}
+{"id":"tinygrad","name":"tinygrad","vendor":"tinycorp / George Hotz","category":"training","language":"Python","license":"MIT","starsK":33.6,"version":"0.14.x","released":"2020-11","features":["minimal ML framework","multiple accelerators","lazy compute graph"],"url":"https://tinygrad.org","github":"https://github.com/tinygrad/tinygrad","notes":"Minimalist deep learning framework. Backs Tinybox products (consumer-grade compute boxes)."}

2026-10-12/podcasts.jsonl ADDED Viewed

	@@ -0,0 +1,50 @@

+{"id":"jkwp2s","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"How to Go from AI Power User to AI Leader","description":"As AI becomes central to work, a new leadership opportunity is emerging for those who can help others use it effectively. In this Operator's Cut, NLW and Nufar explore how AI power users can become leaders on their teams, the skills that distinguish great AI champions, and how companies can build...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/How-to-Go-from-AI-Power-User-to-AI-Leader-e3qcev2","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/127334818/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-12%2F433980474-44100-2-50ecff4b40daf.mp3","duration":"00:39:29","publishedAt":"2026-10-12T14:59:24.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"s35nmk","podcastName":"a16z Podcast","podcastImage":"https://image.simplecastcdn.com/images/0d97354a-306b-45f5-bf26-a8d81eef47ec/ed2664df-9371-438e-8baf-dd2ee0fdde87/3000x3000/thea16zshow-podcastcoverart-3000x3000.jpg?aid=rss_feed","title":"The Race to Recursive Self-Improving AI","description":"a16z General Partner Martin Casado sits down with MakerMaker.AI co-founders Dhaval Adjodah and Owen He to unpack recursive self-improvement, or RSI, and separate the increasingly popular term from what it would actually mean for an AI system to improve itself. Dhaval and Owen distinguish true RSI...","url":"https://a16z.simplecast.com/episodes/the-race-to-recursive-self-improving-ai-OPuX1TcP","audioUrl":"https://mgln.ai/e/1344/afp-848985-injected.calisto.simplecastaudio.com/3f86df7b-51c6-4101-88a2-550dba782de8/episodes/082b16db-6860-47bf-83ca-c3bfaeb02f49/audio/128/default.mp3?aid=rss_feed&amp;awCollectionId=3f86df7b-51c6-4101-88a2-550dba782de8&amp;awEpisodeId=082b16db-6860-47bf-83ca-c3bfaeb02f49&amp;feed=JGE3yC0V","duration":"00:33:27","publishedAt":"2026-10-12T10:00:00.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"24mnm8","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"The 5 Debates Shaping AI","description":"Can AI revenue justify the infrastructure spending? Is AI a mass market or a power-user market? Do businesses want AI they control or simply cheaper models? Who should regulate it? And can data centers win over their neighbors? NLW revisits the format of his most popular episode to explore the five...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/The-5-Debates-Shaping-AI-e3qav4d","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/127285837/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-11%2F433914144-44100-2-0974aebfc49d6.mp3","duration":"00:25:26","publishedAt":"2026-10-11T10:58:13.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"8djq6t","podcastName":"Decoder with Nilay Patel","podcastImage":"https://megaphone.imgix.net/podcasts/f02e8956-cfc6-11f0-83c5-9f12527bae80/image/78986d63d144eb8d46aa4c32a1ada60f.jpg?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"Rewind: Ronan Farrow on Sam Altman’s ‘unconstrained’ relationship with the truth","description":"Hey everybody, it’s Nilay. We’re off today for the holiday, and we’ll be back this Thursday with new episodes. In the meantime, we wanted to revisit a conversation from earlier this year I had with investigative journalist and New Yorker contributor Ronan Farrow, all about the in-depth feature on...","url":"f02e8956-cfc6-11f0-83c5-9f12527bae80","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/pscrb.fm/rss/p/mgln.ai/e/257/traffic.megaphone.fm/VMP1899118519.mp3","duration":"55:55","publishedAt":"2026-10-12T09:00:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"mnfv6t","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"When AI Solves Your Life’s Work","description":"OpenAI’s latest mathematical results are forcing researchers to confront what happens when AI can tackle problems that once defined entire careers. With hundreds of proposed proofs awaiting scrutiny, the debate spans scientific discovery, professional identity, and which fields could face similar...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/When-AI-Solves-Your-Lifes-Work-e3q8kuq","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/127209882/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-9%2F433817882-44100-2-390ed54c878cd.mp3","duration":"00:25:17","publishedAt":"2026-10-09T21:01:23.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"hjq6gc","podcastName":"a16z Podcast","podcastImage":"https://image.simplecastcdn.com/images/0d97354a-306b-45f5-bf26-a8d81eef47ec/ed2664df-9371-438e-8baf-dd2ee0fdde87/3000x3000/thea16zshow-podcastcoverart-3000x3000.jpg?aid=rss_feed","title":"Nick Bostrom: How Should We Navigate Superintelligence?","description":"Nick Bostrom joins Theo Jaffee and Sofia Puccini on MTS to revisit the ideas about superintelligence and existential risk that he helped bring into the mainstream, now that increasingly capable AI systems actually exist. Nick discusses why reinforcement learning may be pushing models toward more...","url":"https://a16z.simplecast.com/episodes/nick-bostrom-how-should-we-navigate-superintelligence-COJSBboX","audioUrl":"https://mgln.ai/e/1344/afp-848985-injected.calisto.simplecastaudio.com/3f86df7b-51c6-4101-88a2-550dba782de8/episodes/eca4cc70-0fd7-4623-bf6f-be7c72d6ed85/audio/128/default.mp3?aid=rss_feed&amp;awCollectionId=3f86df7b-51c6-4101-88a2-550dba782de8&amp;awEpisodeId=eca4cc70-0fd7-4623-bf6f-be7c72d6ed85&amp;feed=JGE3yC0V","duration":"00:53:56","publishedAt":"2026-10-11T10:00:00.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"bgnjgy","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"The Most Important Trends in New AI Products","description":"ChatGPT’s new intelligent UI, Claude Haiku 5.5, and GrokBot’s embrace of rival models reveal where AI is heading: better interfaces, cheaper intelligence, and more choice. NLW breaks down the announcements and why competition is increasingly about the experience built around the models. In the...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/The-Most-Important-Trends-in-New-AI-Products-e3q6nv0","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/127147424/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-8%2F433738897-44100-2-b4fe71a457af3.mp3","duration":"00:27:53","publishedAt":"2026-10-09T00:00:39.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"a680gc","podcastName":"Cognitive Revolution","podcastImage":"https://megaphone.imgix.net/podcasts/1c903b2a-c495-11f1-85f7-d768d17ef297/image/8c237c19a58638569eb38346330576eb.png?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"How Agents Decide: Goodfire's Eric Bigelow on Critical Tokens, Phase Shifts, & In-Context Learning","description":"Goodfire researcher Eric Bigelow joins the show to investigate how large language models arrive at decisions at the level of mechanistic interpretability. Drawing on his research into forking paths, Eric explains that model reasoning functions as in-context learning from sampled tokens, where...","url":"https://www.cognitiverevolution.ai/how-agents-decide-goodfire-s-eric-bigelow-on-critical-tokens-phase-shifts-in-context-learning/","audioUrl":"https://pdst.fm/e/mgln.ai/e/1113/pscrb.fm/rss/p/traffic.megaphone.fm/RINTP6059186610.mp3","duration":"2:02:05","publishedAt":"2026-10-10T10:47:14.000Z","fetchedAt":"2026-10-12T16:25:59.394Z"}
+{"id":"69ifgd","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"A System for Testing New AI Models","description":"In this Operator’s Cut with Nufar Gaspar learn how to test new AI models against your own tasks, compare outputs, and weigh quality, speed, and cost—a repeatable system for figuring out which models deserve a place in your work.Brought to you by:KPMG – Research from KPMG and the University of Texas...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/A-System-for-Testing-New-AI-Models-e3q3meu","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/127047582/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-7%2F433614575-44100-2-a58c6f4f955.mp3","duration":"00:49:49","publishedAt":"2026-10-07T19:02:35.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"ab4eka","podcastName":"a16z Podcast","podcastImage":"https://image.simplecastcdn.com/images/0d97354a-306b-45f5-bf26-a8d81eef47ec/ed2664df-9371-438e-8baf-dd2ee0fdde87/3000x3000/thea16zshow-podcastcoverart-3000x3000.jpg?aid=rss_feed","title":"What Makes a Great Entrepreneur? | Ben Horowitz on Vintage x Vintage","description":"a16z co-founder and General Partner Ben Horowitz joins James Norman on the inaugural episode of Vintage x Vintage for a conversation spanning entrepreneurship, investing, hip-hop, and what changes when AI makes building dramatically easier. Ben reflects on going from Netscape to becoming a...","url":"https://a16z.simplecast.com/episodes/what-makes-a-great-entrepreneur-ben-horowitz-on-vintage-x-vintage-FiiBzjgE","audioUrl":"https://mgln.ai/e/1344/afp-848985-injected.calisto.simplecastaudio.com/3f86df7b-51c6-4101-88a2-550dba782de8/episodes/e56d3db2-edb1-4f5a-a918-c165823f4160/audio/128/default.mp3?aid=rss_feed&amp;awCollectionId=3f86df7b-51c6-4101-88a2-550dba782de8&amp;awEpisodeId=e56d3db2-edb1-4f5a-a918-c165823f4160&amp;feed=JGE3yC0V","duration":"00:51:37","publishedAt":"2026-10-10T10:00:00.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"heax0d","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"Point-Counterpoint: Consumers Will Never Pay for AI","description":"Nearly 98% of US households don’t pay for an AI subscription. Is that a massive untapped market or a sign that Silicon Valley is building for itself? NLW explores both sides of the consumer AI debate, from power-user spending to entertainment and advertising. In the headlines: Reflection unveils...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/Point-Counterpoint-Consumers-Will-Never-Pay-for-AI-e3q16j9","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126965801/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-6%2F433513722-44100-2-69f494a7c8499.mp3","duration":"00:27:26","publishedAt":"2026-10-06T20:19:59.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"tyaojy","podcastName":"Decoder with Nilay Patel","podcastImage":"https://megaphone.imgix.net/podcasts/8229c57c-b924-11f1-b9e8-7b2af0376c77/image/0a5b60eed7a593a85769560128bdbff1.jpg?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"Another billionaire thinks he can fix Warner Bros.","description":"Acquiring Warner Bros. never goes well for anyone. AOL failed, AT&T failed, and Discovery failed — although it did succeed in flipping the assets to David Ellison’s Skydance. So I’m talking to Peter Kafka, Business Insider chief correspondent and host of Channels, to see if Skydance has a chance at...","url":"8229c57c-b924-11f1-b9e8-7b2af0376c77","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/pscrb.fm/rss/p/mgln.ai/e/257/traffic.megaphone.fm/VMP9038342707.mp3","duration":"45:41","publishedAt":"2026-10-10T09:00:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"ca3hsw","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"Why Companies Want AI They Can Own","description":"Companies increasingly want AI they can customize, control, and run on their own terms. NLW explores how that demand could fuel an American open weight AI resurgence—and where safety and national security priorities collide. In the headlines: Amazon’s data center community pledge, Sam Altman on...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/Why-Companies-Want-AI-They-Can-Own-e3pv4le","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126898286/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-5%2F433426135-44100-2-ac2ef5eb3e061.mp3","duration":"00:26:49","publishedAt":"2026-10-05T20:33:05.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"wn09yd","podcastName":"Machine Learning Street Talk","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_episode/4981699/4981699-1791611193615-5ef182c8cdfd5.jpg","title":"What Most People Get Wrong About Evolution | Akarsh Kumar","description":"A lot of people in AI treat evolution as a dumb fallback, basically random search for when you can't take a gradient. Akarsh Kumar thinks that is wrong. Selection hangs on to partial solutions, so mutations only need to be useful about 1% of the time for the search to keep making progress.Akarsh is...","url":"https://podcasters.spotify.com/pod/show/machinelearningstreettalk/episodes/What-Most-People-Get-Wrong-About-Evolution--Akarsh-Kumar-e3q8ds7","audioUrl":"https://traffic.megaphone.fm/APO3123732451.mp3","duration":"00:47:54","publishedAt":"2026-10-10T05:48:01.000Z","fetchedAt":"2026-10-12T16:26:01.649Z"}
+{"id":"vn6d68","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"How to Choose Your Personal AI Agent","description":"Dots, Muse, GrokBot, OpenClaw and more are competing to become your personal AI agent. But which one is right for you? NLW breaks down the trade-offs that matter most, from work versus personal use to model choice, ease of setup and data privacy, and walks through an interactive quiz designed to...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/How-to-Choose-Your-Personal-AI-Agent-e3prr4p","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126790233/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-4%2F433292526-44100-2-f9872d8da355c.mp3","duration":"00:24:31","publishedAt":"2026-10-04T10:56:13.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"kwuw7a","podcastName":"All-In Podcast","podcastImage":"https://static.libsyn.com/p/assets/0/5/1/0/05106eb3c174bb5de5bbc093207a2619/1_Pod_E292.png","title":"Is Claude Conscious? Pope Rejects, Model Welfare Movement, OpenAI's Math Backlash, France Riots","description":"(0:00) Bestie intros! (2:53) Anthropic believes Claude might be conscious: a new religion? Lobbying the Pope, alignment risks, model welfare (38:28) OpenAI's math breakthroughs and community backlash (1:02:03) Riots in France, Friedberg's socialism metric, bond market (1:21:41) Grok Bot goes...","url":"https://allinchamathjason.libsyn.com/is-claude-conscious-pope-rejects-model-welfare-movement-openais-math-backlash-france-riots","audioUrl":"https://dts.podtrac.com/redirect.mp3/traffic.libsyn.com/secure/allinchamathjason/ALLIN-E292_Ch.mp3?dest-id=1928300","duration":"01:33:15","publishedAt":"2026-10-10T02:06:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"e6on5d","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"What the Best Business AI Users Are Doing Different","description":"What separates businesses getting real returns from AI from those still experimenting? New KPMG research reveals how leading organizations are scaling agents, managing multiple models, and connecting AI spending to business value—with ambitions increasingly extending beyond efficiency to revenue...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/What-the-Best-Business-AI-Users-Are-Doing-Different-e3pp6f5","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126703525/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-2%2F433184971-44100-2-7b93ea853ffb3.mp3","duration":"00:22:53","publishedAt":"2026-10-02T18:18:25.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"yevdxe","podcastName":"Latent Space","podcastImage":"https://substackcdn.com/feed/podcast/1084089/post/219666029/b4e32d437c19ecedf77909d1304fdd93.jpg","title":"Why AlphaFold Didn't Solve Protein Folding — Pushmeet Kohli, Google DeepMind & Sal Candido, Biohub","description":"From the Bitter Lesson of AI scaling to the unsolved mysteries of protein folding, Google DeepMind’s Pushmeet Kohli and Biohub’s Sal Candido are rethinking what it takes to build AI that truly understands biology. In this special panel moderated by Brandon Anderson, they explore why AlphaFold’s...","url":"https://www.latent.space/p/biohub-deepmind","audioUrl":"https://api.substack.com/feed/podcast/219666029/8fefb52482ab31b6be94745a76cff81f.mp3","duration":"31:56","publishedAt":"2026-10-10T00:31:26.000Z","fetchedAt":"2026-10-12T16:25:56.504Z"}
+{"id":"r176j2","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"Gemini 4 Argon, Sonnet 5.5 and What Matters with AI Models","description":"Google’s Gemini 4 Argon impresses on benchmarks, but does it signal a real comeback? NLW examines Gemini 4, Claude Sonnet 5.5, and what Muse versus Dots reveals about choosing AI tools. In the headlines: Trump’s super intelligence accord, the America.gov AI portal, and an FTC investigation into...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/Gemini-4-Argon--Sonnet-5-5-and-What-Matters-with-AI-Models-e3pnj53","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126650979/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-1%2F433115992-44100-2-1af35d9d568da.mp3","duration":"00:29:00","publishedAt":"2026-10-01T20:42:00.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"9sx79w","podcastName":"No Priors","podcastImage":"https://megaphone.imgix.net/podcasts/994123ec-c395-11f1-b01e-fff03490a570/image/a780ca93bbc1d2ef87196858bbf07e31.jpg?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"Beam: The Great American Open Model with ReflectionAI Co-Founder and CEO Misha Laskin","description":"Is the future of frontier AI open or closed? ReflectionAI co-founder and CEO Misha Laskin joins Sarah Guo and Elad Gil to discuss the launch of Beam, the company’s 500 billion parameter open-weight reasoning model. Misha breaks down the pre-training and reinforcement learning required to produce...","url":"994123ec-c395-11f1-b01e-fff03490a570","audioUrl":"https://traffic.megaphone.fm/PDP1028435739.mp3","duration":"1:09:44","publishedAt":"2026-10-09T17:24:00.000Z","fetchedAt":"2026-10-12T16:25:59.394Z"}
+{"id":"1xkhs8","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"The Most Important New AI Tools from OpenAI DevDay","description":"OpenAI unveiled more than 20 launches at Dev Day, from always-on Dots agents and the shared workspace Space to cheaper models and new ways to use your ChatGPT subscription across other apps. NLW breaks down the most important announcements, the early reactions, and what they reveal about AI’s shift...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/The-Most-Important-New-AI-Tools-from-OpenAI-DevDay-e3pl7uv","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126573983/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-30%2F433015204-44100-2-5ae726073b88.mp3","duration":"00:23:29","publishedAt":"2026-09-30T18:42:19.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"1tu5p5","podcastName":"Eye on AI","podcastImage":"https://static.libsyn.com/p/assets/4/2/7/1/4271047a3b8c8d1d88c4a68c3ddbc4f2/CS-390-L_AUDIO_THB1_V1.jpg","title":"Chat GPT's Creator Explains The Shocking Truth About How AI Understands the World | Ilya Sutskever","description":"Craig Smith sits down with Ilya Sutskever - Co-Founder and Chief Scientist at Safe Superintelligence Inc. & former chief scientist at OpenAI and one of the primary minds behind GPT-3, GPT-4, and the deep learning revolution that preceded it - for a conversation that covers the intellectual history...","url":"https://aneyeonai.libsyn.com/chat-gpts-creator-explains-the-shocking-truth-about-how-ai-understands-the-world-ilya-sutskever","audioUrl":"https://dts.podtrac.com/redirect.mp3/pscrb.fm/rss/p/pdst.fm/e/arttrk.com/p/VI4CS/prfx.byspotify.com/e/mgln.ai/e/p58841/pscrb.fm/rss/p/traffic.libsyn.com/secure/aneyeonai/CS-390-LAudio_V2.mp3?dest-id=727317","duration":"41:45","publishedAt":"2026-10-09T15:00:00.000Z","fetchedAt":"2026-10-12T16:26:07.250Z"}
+{"id":"obp12l","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"How to Build Team Agents","description":"Nufar Gaspar joins this AIDB Operator’s Cut to explore how to build AI agents that work for an entire team. The conversation focuses on moving from individual AI use to shared agents that support collaborative work.Register for our Free Webinar: Build Your Personal AI Benchmark -...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/How-to-Build-Team-Agents-e3pj5qj","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126506259/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-29%2F432928767-44100-2-daaa1676841fc.mp3","duration":"00:41:48","publishedAt":"2026-09-29T18:39:47.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"32p954","podcastName":"The Vergecast","podcastImage":"https://megaphone.imgix.net/podcasts/e4c412e6-e1f0-11e8-9bde-83f9d376f059/image/fa367fa8b1b6ca01fc4b8c34d83ba4d8.jpg?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"The AI is in the computer","description":"It's gadget season. More specifically, it's apparently \"new computer for your new AI\" season. David and Nilay talk through a slew of announcements — Amazon's Alexa Tablets, Microsoft's Surface Laptop Ultra, the new Googlebooks, and Apple's upcoming smart home launch — to figure out if anyone has...","url":"cb42cff4-c3c7-11f0-8340-174a62840202","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/pscrb.fm/rss/p/mgln.ai/e/257/traffic.megaphone.fm/VMP8833126862.mp3","duration":"1:45:26","publishedAt":"2026-10-09T13:38:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"f0408j","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"The Real Risks of AI Agents","description":"AI agents don’t have to pose an existential threat to cause serious disruption. NLW examines recent OpenAI security incidents, the limits of agent permissions, and how agents doing exactly what users want could upend systems built around human friction. In the headlines: US-China AI talks, Trump...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/The-Real-Risks-of-AI-Agents-e3pgsvg","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126431664/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-28%2F432833155-44100-2-2d9a2cafc1df1.mp3","duration":"00:29:23","publishedAt":"2026-09-28T17:58:52.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"qef5rc","podcastName":"Hard Fork","podcastImage":"https://image.simplecastcdn.com/images/4105a47a-42e5-4ccc-887a-832af7989986/23965394-f5e4-4fdb-b150-639f4910353e/3000x3000/nyt-hf-album-art-3000-2.jpg?aid=rss_feed","title":"Anthropic’s Quest to Give A.I. Morals","description":"“For Anthropic, it was part research and part evangelism.”","url":"https://www.nytimes.com/column/hard-fork","audioUrl":"https://dts.podtrac.com/redirect.mp3/pdst.fm/e/pfx.vpixl.com/6qj4J/pscrb.fm/rss/p/nyt.simplecastaudio.com/3e43d072-f8a5-430f-bc8e-4c70aafdf3c7/episodes/05ad1bfb-bd47-4120-a5fa-98c2f4787a3d/audio/128/default.mp3?aid=rss_feed&amp;awCollectionId=3e43d072-f8a5-430f-bc8e-4c70aafdf3c7&amp;awEpisodeId=05ad1bfb-bd47-4120-a5fa-98c2f4787a3d&amp;feed=l2i9YnTd","duration":"00:44:20","publishedAt":"2026-10-09T11:00:00.000Z","fetchedAt":"2026-10-12T16:26:01.649Z"}
+{"id":"bcznon","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"The Rise of the AI Moderates","description":"As the AI debate grows more polarized, a growing chorus is rejecting the choice between unchecked optimism and inevitable doom. NLW explores arguments from Francis Fukuyama, AI researchers, and Jeffrey Katzenberg that point toward a more productive conversation about AI risk, human creativity, and...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/The-Rise-of-the-AI-Moderates-e3pet2d","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126366221/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-27%2F432744736-44100-2-0b12dc2a815e.mp3","duration":"00:32:50","publishedAt":"2026-09-27T11:14:50.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"n77ybf","podcastName":"Super Data Science","podcastImage":"https://megaphone.imgix.net/podcasts/436573d4-c308-11f1-881e-b7a095ea3827/image/b25143610712c9bdeac57f63bb5166e1.png?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"1034: In Case You Missed It in September 2026","description":"In ICYMI Episode #1034, Jon Krohn moves from what sits underneath AI systems to what it takes to live with them every day. Hear from Aishwarya Srinivasan, Luis Serrano, Katie Malone, Ish Shah and Dilani Kahawala, discussing why the bottleneck in shipping software has moved from execution to...","url":"436573d4-c308-11f1-881e-b7a095ea3827","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/chrt.fm/track/E581B9/arttrk.com/p/VI4CS/pscrb.fm/rss/p/traffic.megaphone.fm/SUPERDATASCIENCEPTYLTD1161453705.mp3","duration":"43:34","publishedAt":"2026-10-09T11:00:00.000Z","fetchedAt":"2026-10-12T16:26:07.250Z"}
+{"id":"kb6zn0","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"How People Are Actually Using Jev","description":"Jev is moving beyond viral demos into practical work, from analyzing ad campaigns and searching archives to prioritizing inboxes and checking AI writing. NLW breaks down six categories of real-world use cases, explains what makes this new “judgment model” different from an LLM, and offers a...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/How-People-Are-Actually-Using-Jev-e3pd0rj","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126304563/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-25%2F432662503-44100-2-4dbca7b913e13.mp3","duration":"00:25:57","publishedAt":"2026-09-25T18:07:32.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"cbi2no","podcastName":"a16z Podcast","podcastImage":"https://image.simplecastcdn.com/images/0d97354a-306b-45f5-bf26-a8d81eef47ec/ed2664df-9371-438e-8baf-dd2ee0fdde87/3000x3000/thea16zshow-podcastcoverart-3000x3000.jpg?aid=rss_feed","title":"Chips, Memory, and Power | Pat Gelsinger","description":"a16z's Raghu Raghuram and Guido Appenzeller sit down with Playground Global General Partner and former Intel CEO Pat Gelsinger to discuss the next wave of semiconductor innovation and the physical constraints shaping the AI buildout. Drawing on his experience designing Intel's 386 and 486...","url":"https://a16z.simplecast.com/episodes/chips-memory-and-power-pat-gelsinger-uwHe_vsU","audioUrl":"https://mgln.ai/e/1344/afp-848985-injected.calisto.simplecastaudio.com/3f86df7b-51c6-4101-88a2-550dba782de8/episodes/2a98eb9e-80e8-4172-958c-8f92f8976e78/audio/128/default.mp3?aid=rss_feed&amp;awCollectionId=3f86df7b-51c6-4101-88a2-550dba782de8&amp;awEpisodeId=2a98eb9e-80e8-4172-958c-8f92f8976e78&amp;feed=JGE3yC0V","duration":"00:54:15","publishedAt":"2026-10-09T10:00:00.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"kepgcf","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"AI Agents Are Moving Into the Real World","description":"Meta is bringing Muse to smart glasses and a new wearable, while GrokBot is turning Teslas into voice-controlled personal assistants. As skeptics question whether consumers actually want AI agents, early users are finding real value in handing off life's annoying admin. In the headlines: Claude’s...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/AI-Agents-Are-Moving-Into-the-Real-World-e3pbisu","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126257502/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-24%2F432598497-44100-2-61fee1b301566.mp3","duration":"00:28:07","publishedAt":"2026-09-24T19:06:43.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"6t4bj0","podcastName":"Last Week in AI","podcastImage":"https://content.production.cdn.art19.com/images/64/91/19/9f/6491199f-bab4-46cc-a940-2e46a7a9d8aa/79788de05ff297d5e2fade962c95430c8a2d7956ade95cdf4919cfcd3a0f79f7c9a1f4ddb825f37ab6c09445022dcaf8bb07095b999012c322da5e3d49b53266.jpeg","title":"#259 - Dots, Sonnet, Self-Safety, Rogue AI","description":"Our 259th episode with a summary and discussion of last week's big AI news!Recorded on 09/30/2026 ; I lied about this one coming out soon... but not the next one , I promise!Hosted by Andrey Kurenkov and Jeremie HarrisFeel free to email us your questions and feedback at andreyvkurenkov@gmail.com...","url":"gid://art19-episode-locator/V0/BiNXcdDi83I9vwlOsfBeO5b_B8TjjipW6NpeGlXCSvA","audioUrl":"https://mgln.ai/e/211/rss.art19.com/episodes/4cbdcb0f-42f4-4952-bcce-f19cb787da89.mp3?rss_browser=BAhJIg9UZW5zb3JGZWVkBjoGRVQ%3D--b5747a94e27ba1ca6ee95433d2182d7c7cfcd74f","duration":"01:25:20","publishedAt":"2026-10-09T07:00:00.000Z","fetchedAt":"2026-10-12T16:25:59.394Z"}
+{"id":"mo7yfl","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"Opus 5.5 vs GPT-6 Sol and Luna","description":"Anthropic and OpenAI release new models on the same day, with Opus 5.5 winning early acclaim while GPT-6 Sol and Luna push the frontier of affordable intelligence. NLW breaks down the benchmarks, real-world use cases, and first reactions—and why model personality, falling costs, and the tools...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/Opus-5-5-vs-GPT-6-Sol-and-Luna-e3pa5c4","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126210884/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-23%2F432533842-44100-2-99559049bf5a5.mp3","duration":"00:31:02","publishedAt":"2026-09-23T20:50:01.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"sm1z33","podcastName":"The Vergecast","podcastImage":"https://megaphone.imgix.net/podcasts/e4c412e6-e1f0-11e8-9bde-83f9d376f059/image/fa367fa8b1b6ca01fc4b8c34d83ba4d8.jpg?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"The many pluses and minuses of Alexa Plus","description":"Amazon is determined to turn Alexa Plus into a do-everything AI assistant that can compete with Gemini, ChatGPT, Claude, and the rest. But after a year of testing, The Verge's Jen Tuohy has mixed feelings on how it's all going. And on whether a new tablet is the right way to fix it. Further...","url":"4a1922ac-6e7b-11f1-94bf-3f97b2f47f48","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/pscrb.fm/rss/p/mgln.ai/e/257/traffic.megaphone.fm/VMP8056613997.mp3","duration":"34:51","publishedAt":"2026-10-08T19:22:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"br2ufp","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"Agent Wars!","description":"Meta’s Muse has overtaken ChatGPT in the App Store, and Amazon has responded by blocking it from shopping on its site. Shopify is opening its doors instead, setting up a fight over who owns the customer relationship when personal agents make decisions and purchases. In the headlines: Grok 4.7, AI...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/Agent-Wars-e3p8f3d","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126155309/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-22%2F432457268-44100-2-e0a7717f54c4e.mp3","duration":"00:30:57","publishedAt":"2026-09-22T19:07:36.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"boa2kw","podcastName":"Latent Space","podcastImage":"https://substackcdn.com/feed/podcast/1084089/post/219146438/97b1084778de9df0ba80e614d6a0f1c9.jpg","title":"Synthesis Superintelligence: from Semiconductors to Superconductors — Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk","description":"It’s hard to believe that Periodic was only launched last September:One year later, it is considered one of the pre-eminent AI scientist labs, with dizzying talent density and astonishing progress in the autonomous lab buildout:Most people are familiar with the standard credentials of Liam and...","url":"https://www.latent.space/p/periodic","audioUrl":"https://api.substack.com/feed/podcast/219146438/4317811e52984f32f5f67455cb91c1e9.mp3","duration":"1:24:00","publishedAt":"2026-10-08T16:27:54.000Z","fetchedAt":"2026-10-12T16:25:56.504Z"}
+{"id":"1bs5ep","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"The State of the AI Debate","description":"Trump wants an AI Force, politicians are pushing kill switches, and calls for a slowdown keep running into the same question about China. Ahead of this week’s Trump–Xi meeting, NLW explores the competing priorities shaping AI policy and the prospects for cooperation between the two AI superpowers....","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/The-State-of-the-AI-Debate-e3p6r0p","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126101977/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-21%2F432386739-44100-2-12f7014a45b6c.mp3","duration":"00:27:12","publishedAt":"2026-09-21T20:40:10.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"2z90x0","podcastName":"The MAD Podcast","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/40657026/40657026-1768486187830-f8f544fbc7c46.jpg","title":"What Happens When Billions of AI Agents Hit Your Database? (Andy Pavlo)","description":"AI agents are becoming the biggest users of databases: creating them, querying them, and sometimes deleting them. Andy Pavlo, Carnegie Mellon's database professor and now VP of Database Research at ClickHouse, explains what changes when billions of agents hit the data layer, and why most of the AI...","url":"https://podcasters.spotify.com/pod/show/firstmark/episodes/What-Happens-When-Billions-of-AI-Agents-Hit-Your-Database--Andy-Pavlo-e3q4qu7","audioUrl":"https://anchor.fm/s/f2ee4948/podcast/play/127084935/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-9-8%2F433665610-44100-2-a2e92ad5e51b2.mp3","duration":"01:16:46","publishedAt":"2026-10-08T11:30:00.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"fp31yg","podcastName":"AI Daily Brief","podcastImage":"https://d3t3ozftmdmh3i.cloudfront.net/staging/podcast_uploaded_nologo/41472609/41472609-1752234663609-8665756a468e5.jpg","title":"7 Ways How We Use AI Is Changing","description":"From persistent conversations and voice commands to chatbots that coordinate entire teams of agents, the way we work with AI is shifting. NLW breaks down seven changes shaping everyday AI use, including the move from prompts to goals, managing model costs, and building shared agents for...","url":"https://podcasters.spotify.com/pod/show/nlw/episodes/7-Ways-How-We-Use-AI-Is-Changing-e3p4fqt","audioUrl":"https://anchor.fm/s/f7cac464/podcast/play/126024989/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-8-20%2F432284493-44100-2-de745c4e4f594.mp3","duration":"00:26:21","publishedAt":"2026-09-20T11:28:50.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"3q0p2n","podcastName":"Cognitive Revolution","podcastImage":"https://megaphone.imgix.net/podcasts/5d1fb616-c2fb-11f1-bafc-cff9298597f8/image/aef7552479d90fcc19ab25212ac884c5.png?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"AI:AM: A Level We Shouldn't Pass? Notes from The Curve + Tokens vs. Salaries & Is SaaS Cooked?","description":"Nathan Labenz reports back from The Curve conference, sharing off-the-record insights from frontier lab leaders on shortened timelines and whether humanity is approaching an intelligence threshold it should not cross. The episode also features discussions with Positron CTO Thomas Sohmers on...","url":"https://www.cognitiverevolution.ai/ai-am-a-level-we-shouldn-t-pass-notes-from-the-curve-tokens-vs-salaries-is-saas-cooked/","audioUrl":"https://pdst.fm/e/mgln.ai/e/1113/pscrb.fm/rss/p/traffic.megaphone.fm/RINTP1992288352.mp3","duration":"1:28:00","publishedAt":"2026-10-08T10:50:43.000Z","fetchedAt":"2026-10-12T16:25:59.394Z"}
+{"id":"t3zknq","podcastName":"a16z Podcast","podcastImage":"https://image.simplecastcdn.com/images/0d97354a-306b-45f5-bf26-a8d81eef47ec/ed2664df-9371-438e-8baf-dd2ee0fdde87/3000x3000/thea16zshow-podcastcoverart-3000x3000.jpg?aid=rss_feed","title":"Building the Cloud for an Agentic World | AWS CEO Matt Garman","description":"a16z’s Raghu Raghuram sits down with AWS CEO Matt Garman to discuss how AI is reshaping the cloud, from the needs of AI-native startups to infrastructure increasingly designed for agents. Matt explains how AWS is adapting as agents write code and manage infrastructure, why it’s reserving scarce GPU...","url":"https://a16z.simplecast.com/episodes/building-the-cloud-for-an-agentic-world-aws-ceo-matt-garman-d18BWOHC","audioUrl":"https://mgln.ai/e/1344/afp-848985-injected.calisto.simplecastaudio.com/3f86df7b-51c6-4101-88a2-550dba782de8/episodes/e1b4fcb1-9600-4c7b-a712-186a9c63f2fd/audio/128/default.mp3?aid=rss_feed&amp;awCollectionId=3f86df7b-51c6-4101-88a2-550dba782de8&amp;awEpisodeId=e1b4fcb1-9600-4c7b-a712-186a9c63f2fd&amp;feed=JGE3yC0V","duration":"00:56:17","publishedAt":"2026-10-08T10:00:00.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"gic3dc","podcastName":"Practical AI","podcastImage":"https://img.transistorcdn.com/ainvSatjVoL-Lmws17e1DNFjXP2EryjtfMxrjh18hpc/rs:fill:0:0:1/w:1400/h:1400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS8yYTcx/YzU2YWUwYTA1MDVl/YTBiOTFmNjYwMWQ0/ZjA4Yy5wbmc.jpg","title":"Narrative intelligence and the human advantage","description":"As AI reshapes how we work, create, and think about our roles, staying adaptable means understanding not just the technology, but our own stories and identities. Reed Frerichs joins Daniel and Chris to explore narrative intelligence, storytelling, leadership, and personal growth in an AI-driven...","url":"https://share.transistor.fm/s/f2fcb677","audioUrl":"https://pscrb.fm/rss/p/dts.podtrac.com/redirect.mp3/media.transistor.fm/f2fcb677/ee0419b3.mp3","duration":"44:24","publishedAt":"2026-10-08T09:00:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"77vis6","podcastName":"Decoder with Nilay Patel","podcastImage":"https://megaphone.imgix.net/podcasts/f0bb4cba-cfc6-11f0-83c5-b30efc51f01d/image/c2c8bd3c4229763664e2a3c800e72fff.jpg?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"OpenAI has a Muse problem","description":"My guest today is Hayden Field, The Verge’s senior AI reporter, and we're discussing the new wave of consumer friendly AI agents. AI enthusiasts have been using agents for a minute now, but the launch of Meta’s Muse, OpenAI’s Dots, and XAI’s Grok Bot has brought easy to use agents to...","url":"f0bb4cba-cfc6-11f0-83c5-b30efc51f01d","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/pscrb.fm/rss/p/mgln.ai/e/257/traffic.megaphone.fm/VMP6813236420.mp3","duration":"45:29","publishedAt":"2026-10-08T09:00:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"gm09fu","podcastName":"The Vergecast","podcastImage":"https://megaphone.imgix.net/podcasts/e4c412e6-e1f0-11e8-9bde-83f9d376f059/image/fa367fa8b1b6ca01fc4b8c34d83ba4d8.jpg?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"The hot new thing in headphones","description":"Bose, Sony, Beats, and seemingly every other headphone maker has recently released a new and excellent set of cans. Or buds. Or both. The Verge's John Higgins explains why the headphone market is suddenly so great, how the companies are competing now, and which new pair is the right one for you....","url":"49fcf4c4-6e7b-11f1-94bf-5737d233658f","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/pscrb.fm/rss/p/mgln.ai/e/257/traffic.megaphone.fm/VMP2488237973.mp3","duration":"40:50","publishedAt":"2026-10-07T19:12:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"javbz8","podcastName":"Cognitive Revolution","podcastImage":"https://megaphone.imgix.net/podcasts/dfa41bc0-c208-11f1-a559-176e7674ab8f/image/4ab05942b1cb934e623b613bb6a55bd6.png?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"Software That Never Breaks: OutSystems CEO Woodson Martin on Building Enterprise-Grade Apps at ...","description":"OutSystems CEO Woodson Martin joins Nathan to discuss how enterprise software can evolve at AI speed without breaking critical operations. Martin explains how an intermediate layer of abstract modeling allows AI agents to build reliably through deterministic code generation, ensuring role-based...","url":"https://www.cognitiverevolution.ai/software-that-never-breaks-outsystems-ceo-woodson-martin-on-building-enterprise-grade-apps-at-ai-speed/","audioUrl":"https://pdst.fm/e/mgln.ai/e/1113/pscrb.fm/rss/p/traffic.megaphone.fm/RINTP9070453666.mp3","duration":"1:11:26","publishedAt":"2026-10-07T17:00:00.000Z","fetchedAt":"2026-10-12T16:25:59.394Z"}
+{"id":"k7o0r4","podcastName":"a16z Podcast","podcastImage":"https://image.simplecastcdn.com/images/0d97354a-306b-45f5-bf26-a8d81eef47ec/ed2664df-9371-438e-8baf-dd2ee0fdde87/3000x3000/thea16zshow-podcastcoverart-3000x3000.jpg?aid=rss_feed","title":"How Valon Rebuilt a $13 Trillion Industry From Scratch","description":"a16z General Partner Angela Strange sits down with Valon’s Andrew Wang and Linda Du to unpack what it takes to rebuild the infrastructure underneath a $13 trillion mortgage market that still relies heavily on systems designed before the internet. Linda and Andrew explain why Valon chose the hardest...","url":"https://a16z.simplecast.com/episodes/how-valon-rebuilt-a-13-trillion-industry-from-scratch-0mcVWvYe","audioUrl":"https://mgln.ai/e/1344/afp-848985-injected.calisto.simplecastaudio.com/3f86df7b-51c6-4101-88a2-550dba782de8/episodes/b653e837-66de-4c84-8582-adba2b75051f/audio/128/default.mp3?aid=rss_feed&amp;awCollectionId=3f86df7b-51c6-4101-88a2-550dba782de8&amp;awEpisodeId=b653e837-66de-4c84-8582-adba2b75051f&amp;feed=JGE3yC0V","duration":"00:40:33","publishedAt":"2026-10-07T10:00:00.000Z","fetchedAt":"2026-10-12T16:26:02.236Z"}
+{"id":"non2gu","podcastName":"TWIML AI","podcastImage":"https://megaphone.imgix.net/podcasts/edcb2910-c1cb-11f1-b902-a7ba02688fa0/image/845752c45a90230ad7dec9740fdb341b.png?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"Why Jev Is Changing How We Build With AI with Diogo Almeida - #779","description":"In this episode, Diogo Almeida, co-founder and CEO of TypeSafe, joins us to discuss Jev, TypeSafe’s recently released model for bringing fast, reliable intelligence directly into software. We explore the idea of “machine-native intelligence” and why Diogo believes models optimized for generating...","url":"https://twimlai.com/podcast/twimlai/jev-changing-how-we-build-ai","audioUrl":"https://pscrb.fm/rss/p/traffic.megaphone.fm/MLN2040611678.mp3","duration":"1:30:09","publishedAt":"2026-10-06T21:43:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"max23m","podcastName":"The Vergecast","podcastImage":"https://megaphone.imgix.net/podcasts/e4c412e6-e1f0-11e8-9bde-83f9d376f059/image/fa367fa8b1b6ca01fc4b8c34d83ba4d8.jpg?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"How much should you tell your AI agent?","description":"The next big thing in AI, at least if you ask the tech companies, is assistant agents. David and The Verge's Allison Johnson have both been testing products like Meta's Muse, OpenAI's Dots, and startups like Instinct, in an effort to find out how much you can really get done by messaging with an AI...","url":"ca7317c8-c3c7-11f0-8340-171c0dbf8e62","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/pscrb.fm/rss/p/mgln.ai/e/257/traffic.megaphone.fm/VMP4220452170.mp3","duration":"41:20","publishedAt":"2026-10-06T19:27:00.000Z","fetchedAt":"2026-10-12T16:26:04.035Z"}
+{"id":"64so0y","podcastName":"Eye on AI","podcastImage":"https://static.libsyn.com/p/assets/3/0/b/8/30b8d57c8e86564ce5bbc093207a2619/CS-388-L_AUDIO_THB1_V2.jpg","title":"The Coordination Tax: Why AI Is Burning Out Your Best People | Dan O'Connell, Front","description":"Everyone assumes AI in customer operations means fewer people and lower costs. The data from 700 customer operations leaders is more complicated, and more honest. Dan O'Connell, CEO of Front, joins Craig Smith with findings from a survey of customer support, service, and account management leaders...","url":"https://aneyeonai.libsyn.com/the-coordination-tax-why-ai-is-burning-out-your-best-people-dan-oconnell-front","audioUrl":"https://dts.podtrac.com/redirect.mp3/pscrb.fm/rss/p/pdst.fm/e/arttrk.com/p/VI4CS/prfx.byspotify.com/e/mgln.ai/e/p58841/pscrb.fm/rss/p/traffic.libsyn.com/secure/aneyeonai/CS-388-LAudio_V2.mp3?dest-id=727317","duration":"51:14","publishedAt":"2026-10-06T14:28:00.000Z","fetchedAt":"2026-10-12T16:26:07.250Z"}
+{"id":"dsajvf","podcastName":"Super Data Science","podcastImage":"https://megaphone.imgix.net/podcasts/9764d488-c098-11f1-989a-37410cdce8cb/image/8008a3ce23c01c5f4e36d482a7e21a86.png?ixlib=rails-4.3.1&amp;max-w=3000&amp;max-h=3000&amp;fit=crop&amp;auto=format,compress","title":"1033: Workslop: The Hidden Cost of AI-Generated Work, with Prof. Jeff Hancock and Dr. Kate Niederhoffer","description":"In Episode #1033, Prof. Jeff Hancock (Professor of Communication at Stanford) and Dr. Kate Niederhoffer (Chief Scientist at BetterUp) join Jon Krohn to explain the hidden cost of AI-generated work. A year ago they coined \"workslop\" in a Harvard Business Review article that went viral and landed the...","url":"9764d488-c098-11f1-989a-37410cdce8cb","audioUrl":"https://www.podtrac.com/pts/redirect.mp3/chrt.fm/track/E581B9/arttrk.com/p/VI4CS/pscrb.fm/rss/p/traffic.megaphone.fm/SUPERDATASCIENCEPTYLTD5656400315.mp3","duration":"1:15:25","publishedAt":"2026-10-06T11:00:00.000Z","fetchedAt":"2026-10-12T16:26:07.250Z"}

2026-10-12/pricing.jsonl ADDED Viewed

	@@ -0,0 +1 @@

+ {"ok":true,"source":"tensorfeed.ai","updated":"2026-10-12T17:17:09.896Z","providers":{"lastUpdated":"2026-10-01","providers":[{"id":"anthropic","name":"Anthropic","logo":"/images/providers/anthropic.png","url":"https://www.anthropic.com","models":[{"id":"claude-opus-5-5","name":"Claude Opus 5.5","inputPrice":4,"outputPrice":20,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"claude-opus-5","name":"Claude Opus 5","inputPrice":5,"outputPrice":25,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"claude-fable-5-1","name":"Claude Fable 5.1","inputPrice":10,"outputPrice":50,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"claude-mythos-5-1","name":"Claude Mythos 5.1","inputPrice":10,"outputPrice":50,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"claude-fable-5","name":"Claude Fable 5","inputPrice":10,"outputPrice":50,"contextWindow":1000000,"released":"2026-06","capabilities":["text","vision","tool-use","code"],"tier":"flagship"},{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","inputPrice":2,"outputPrice":10,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"claude-sonnet-5","name":"Claude Sonnet 5","inputPrice":2,"outputPrice":10,"contextWindow":1000000,"released":"2026-06","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"claude-opus-4-8","name":"Claude Opus 4.8","inputPrice":5,"outputPrice":25,"contextWindow":1000000,"released":"2026-05","capabilities":["text","vision","tool-use","code"],"tier":"flagship"},{"id":"claude-opus-4-7","name":"Claude Opus 4.7","inputPrice":5,"outputPrice":25,"contextWindow":1000000,"released":"2026-04","capabilities":["text","vision","tool-use","code"],"tier":"flagship"},{"id":"claude-opus-4-6","name":"Claude Opus 4.6","inputPrice":5,"outputPrice":25,"contextWindow":1000000,"released":"2026-03","capabilities":["text","vision","tool-use","code"],"tier":"mid"},{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","inputPrice":3,"outputPrice":15,"contextWindow":1000000,"released":"2026-03","capabilities":["text","vision","tool-use","code"],"tier":"mid"},{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","inputPrice":1,"outputPrice":5,"contextWindow":200000,"released":"2025-10","capabilities":["text","vision","tool-use","code"],"tier":"budget"}]},{"id":"openai","name":"OpenAI","logo":"/images/providers/openai.png","url":"https://openai.com","models":[{"id":"gpt-6-astra","name":"GPT-6 Astra","inputPrice":10,"outputPrice":50,"contextWindow":1050000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"gpt-6-1-sol","name":"GPT-6.1 Sol","inputPrice":2,"outputPrice":10,"contextWindow":1050000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"gpt-6-sol","name":"GPT-6 Sol","inputPrice":2,"outputPrice":10,"contextWindow":1050000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"gpt-6-luna","name":"GPT-6 Luna","inputPrice":0.1,"outputPrice":0.5,"contextWindow":1050000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget"},{"id":"gpt-5-6-sol","name":"GPT-5.6 Sol","inputPrice":4,"outputPrice":20,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"gpt-5-6-terra","name":"GPT-5.6 Terra","inputPrice":2,"outputPrice":12,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"gpt-5-6-luna","name":"GPT-5.6 Luna","inputPrice":0.2,"outputPrice":1.2,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget"},{"id":"gpt-5-5","name":"GPT-5.5","inputPrice":5,"outputPrice":30,"contextWindow":1000000,"released":"2026-04","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"gpt-4o","name":"GPT-4o","inputPrice":2.5,"outputPrice":10,"contextWindow":128000,"released":"2024-05","capabilities":["text","vision","tool-use","code"],"tier":"flagship"},{"id":"gpt-4o-mini","name":"GPT-4o-mini","inputPrice":0.15,"outputPrice":0.6,"contextWindow":128000,"released":"2024-07","capabilities":["text","vision","tool-use","code"],"tier":"budget"},{"id":"o1","name":"o1","inputPrice":15,"outputPrice":60,"contextWindow":200000,"released":"2024-12","capabilities":["text","reasoning","code"],"tier":"flagship"},{"id":"o3-mini","name":"o3-mini","inputPrice":1.1,"outputPrice":4.4,"contextWindow":200000,"released":"2025-01","capabilities":["text","reasoning","code"],"tier":"mid"}]},{"id":"google","name":"Google","logo":"/images/providers/google.png","url":"https://ai.google.dev","models":[{"id":"gemini-2-5-pro","name":"Gemini 2.5 Pro","inputPrice":1.25,"outputPrice":10,"contextWindow":1048576,"released":"2025-03","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"gemini-3-1-flash-lite","name":"Gemini 3.1 Flash-Lite","inputPrice":0.25,"outputPrice":1.5,"contextWindow":1048576,"released":"2026-05","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget"},{"id":"gemini-3-5-flash","name":"Gemini 3.5 Flash","inputPrice":1.5,"outputPrice":9,"contextWindow":1048576,"released":"2026-05","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"gemini-3-8-flash","name":"Gemini 3.8 Flash","inputPrice":0.75,"outputPrice":3.75,"contextWindow":1048576,"released":"2026-09","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"mid"},{"id":"gemini-3-7-flash","name":"Gemini 3.7 Flash","inputPrice":0.75,"outputPrice":3.75,"contextWindow":1048576,"released":"2026-08","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"mid"},{"id":"gemini-3-6-flash","name":"Gemini 3.6 Flash","inputPrice":0.75,"outputPrice":3.75,"contextWindow":1048576,"released":"2026-07","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"mid"},{"id":"gemini-3-5-flash-lite","name":"Gemini 3.5 Flash-Lite","inputPrice":0.3,"outputPrice":2.5,"contextWindow":1048576,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget"}]},{"id":"meta","name":"Meta","logo":"/images/providers/meta.png","url":"https://ai.meta.com","models":[{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","inputPrice":0.35,"outputPrice":1.5,"contextWindow":131072,"released":"2026-08","openSource":true,"license":"Apache-2.0","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"muse-spark-1-3","name":"Muse Spark 1.3","inputPrice":1.25,"outputPrice":4.25,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"muse-spark-1-3-contributor","name":"Muse Spark 1.3 Contributor","inputPrice":0.1,"outputPrice":0.2,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget"},{"id":"muse-spark-1-2","name":"Muse Spark 1.2","inputPrice":1.25,"outputPrice":4.25,"contextWindow":1000000,"released":"2026-08","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"muse-spark-1-2-contributor","name":"Muse Spark 1.2 Contributor","inputPrice":0.1,"outputPrice":0.2,"contextWindow":1000000,"released":"2026-08","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"budget"},{"id":"muse-spark-1-1","name":"Muse Spark 1.1","inputPrice":1.25,"outputPrice":4.25,"contextWindow":1000000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"llama-4-scout","name":"Llama 4 Scout","inputPrice":0,"outputPrice":0,"contextWindow":10000000,"released":"2025-04","openSource":true,"license":"Llama 4 Community License","capabilities":["text","vision","code"],"tier":"mid"},{"id":"llama-4-maverick","name":"Llama 4 Maverick","inputPrice":0,"outputPrice":0,"contextWindow":1000000,"released":"2025-04","openSource":true,"license":"Llama 4 Community License","capabilities":["text","vision","code"],"tier":"flagship"}]},{"id":"mistral","name":"Mistral","logo":"/images/providers/mistral.png","url":"https://mistral.ai","models":[{"id":"mistral-medium-3-5","name":"Mistral Medium 3.5","inputPrice":1.5,"outputPrice":7.5,"contextWindow":256000,"released":"2026-04","openSource":true,"license":"Modified MIT","capabilities":["text","code","tool-use","reasoning"],"tier":"mid"},{"id":"mistral-large","name":"Mistral Large","inputPrice":0.5,"outputPrice":1.5,"contextWindow":262144,"released":"2025-12","openSource":true,"license":"Apache-2.0","capabilities":["text","vision","tool-use","code"],"tier":"flagship"},{"id":"mistral-small","name":"Mistral Small","inputPrice":0.15,"outputPrice":0.6,"contextWindow":262144,"released":"2026-03","openSource":true,"license":"Apache-2.0","capabilities":["text","tool-use","code"],"tier":"budget"}]},{"id":"cohere","name":"Cohere","logo":"/images/providers/cohere.png","url":"https://cohere.com","models":[{"id":"command-r-plus","name":"Command R+","inputPrice":2.5,"outputPrice":10,"contextWindow":128000,"released":"2024-04","capabilities":["text","tool-use","RAG"],"tier":"flagship"},{"id":"command-r","name":"Command R","inputPrice":0.5,"outputPrice":1.5,"contextWindow":128000,"released":"2024-03","capabilities":["text","tool-use","RAG"],"tier":"mid"}]},{"id":"deepseek","name":"DeepSeek","logo":"/images/providers/deepseek.png","url":"https://www.deepseek.com","models":[{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","inputPrice":1.32,"outputPrice":3.96,"contextWindow":1048576,"released":"2026-04","openSource":true,"license":"MIT","capabilities":["text","tool-use","code","reasoning"],"tier":"flagship"},{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","inputPrice":0.3,"outputPrice":1.2,"contextWindow":1048576,"released":"2026-09","openSource":true,"license":"MIT","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"}]},{"id":"alibaba","name":"Alibaba","logo":"/images/providers/alibaba.png","url":"https://qwenlm.ai","models":[{"id":"qwen3-8-27b","name":"Qwen3.8 27B","inputPrice":0.5,"outputPrice":3,"contextWindow":1000000,"released":"2026-08","openSource":true,"license":"Apache-2.0","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"mid"},{"id":"qwen3-8-a95b","name":"Qwen3.8 2.4T-A95B","inputPrice":2,"outputPrice":6,"contextWindow":1048576,"released":"2026-08","openSource":true,"license":"Qwen3.8-Max License","capabilities":["text","tool-use","code","reasoning"],"tier":"flagship"},{"id":"qwen3-8-max","name":"Qwen3.8-Max","inputPrice":2,"outputPrice":6,"contextWindow":1000000,"released":"2026-08","capabilities":["text","vision","video","code","reasoning","tool-use"],"tier":"flagship"},{"id":"qwen3-8-max-prime","name":"Qwen3.8 Max Prime","inputPrice":4,"outputPrice":12,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","video","code","reasoning","tool-use"],"tier":"flagship"},{"id":"qwen3-7-max","name":"Qwen3.7-Max","inputPrice":2.5,"outputPrice":7.5,"contextWindow":1000000,"released":"2026-05","capabilities":["text","code","reasoning","tool-use"],"tier":"flagship"},{"id":"qwen3-8-omni-flash","name":"Qwen3.8-Omni-Flash","inputPrice":0.15,"outputPrice":0.47,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"budget"},{"id":"qwen3-8-flash","name":"Qwen3.8-Flash","inputPrice":0.15,"outputPrice":0.47,"contextWindow":1000000,"released":"2026-08","capabilities":["text","vision","video","tool-use","code","reasoning"],"tier":"budget"}]},{"id":"xai","name":"xAI","logo":"/images/providers/xai.png","url":"https://x.ai","models":[{"id":"grok-4-7","name":"Grok 4.7","inputPrice":2,"outputPrice":6,"contextWindow":500000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"grok-4-6","name":"Grok 4.6","inputPrice":2,"outputPrice":6,"contextWindow":500000,"released":"2026-08","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"grok-4-5","name":"Grok 4.5","inputPrice":2,"outputPrice":6,"contextWindow":500000,"released":"2026-07","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"grok-4-3","name":"Grok 4.3","inputPrice":1.25,"outputPrice":2.5,"contextWindow":1000000,"released":"2026-04","capabilities":["text","vision","tool-use","code"],"tier":"flagship"}]},{"id":"nvidia","name":"NVIDIA","logo":"/images/providers/nvidia.png","url":"https://www.nvidia.com/en-us/ai/","models":[{"id":"nemotron-3-5-lightning","name":"Nemotron 3.5 Lightning","inputPrice":0.08,"outputPrice":0.2,"contextWindow":262144,"released":"2026-08","openSource":true,"license":"OpenMDW-1.1","capabilities":["text","tool-use","code","reasoning"],"tier":"budget"},{"id":"nemotron-3-nano-omni","name":"Nemotron 3 Nano Omni","inputPrice":0,"outputPrice":0,"contextWindow":256000,"released":"2026-04","openSource":true,"license":"NVIDIA Open Model License","capabilities":["text","vision","audio","video","code","reasoning","tool-use"],"tier":"mid"}]},{"id":"microsoft","name":"Microsoft","logo":"/images/providers/microsoft.png","url":"https://microsoft.ai","models":[{"id":"mai-code-1-1-flash","name":"MAI-Code-1.1-Flash","inputPrice":0.2,"outputPrice":1.2,"contextWindow":256000,"released":"2026-08","capabilities":["text","vision","code","tool-use"],"tier":"budget"}]},{"id":"minimax","name":"MiniMax","logo":"/images/providers/minimax.png","url":"https://www.minimax.io","models":[{"id":"minimax-m3","name":"MiniMax M3","inputPrice":0.3,"outputPrice":1.2,"contextWindow":1048576,"released":"2026-06","openSource":true,"license":"MiniMax Community License","capabilities":["text","vision","video","code","tool-use"],"tier":"budget"}]},{"id":"meituan","name":"Meituan","logo":"/images/providers/meituan.png","url":"https://www.longcatai.org","models":[{"id":"longcat-2","name":"LongCat-2.0","inputPrice":0.75,"outputPrice":2.95,"contextWindow":1000000,"released":"2026-06","openSource":true,"license":"MIT","capabilities":["text","code","tool-use","reasoning"],"tier":"flagship"},{"id":"longcat-2-5-preview","name":"LongCat-2.5-Preview","inputPrice":0.3,"outputPrice":1.2,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"}]},{"id":"zhipu","name":"Z.ai (Zhipu AI)","logo":"/images/providers/zhipu.png","url":"https://z.ai","models":[{"id":"glm-5-3","name":"GLM-5.3","inputPrice":1.4,"outputPrice":4.4,"contextWindow":1048576,"released":"2026-08","openSource":true,"license":"GLM-5.3 License","capabilities":["text","tool-use","code","reasoning"],"tier":"flagship"},{"id":"glm-5-2","name":"GLM-5.2","inputPrice":1.4,"outputPrice":4.4,"contextWindow":1000000,"released":"2026-06","openSource":true,"license":"MIT","capabilities":["text","code","tool-use","reasoning"],"tier":"flagship"},{"id":"glm-5-3-flashx","name":"GLM-5.3-FlashX","inputPrice":0.37,"outputPrice":1.25,"contextWindow":1048576,"released":"2026-09","capabilities":["text","vision","video","tool-use","code","reasoning"],"tier":"budget"},{"id":"glm-5-3-flash","name":"GLM-5.3 Flash","inputPrice":0.15,"outputPrice":0.5,"contextWindow":1048576,"released":"2026-08","openSource":true,"license":"MIT","capabilities":["text","vision","video","tool-use","code","reasoning"],"tier":"budget"}]},{"id":"moonshot","name":"Moonshot AI","logo":"/images/providers/moonshot.png","url":"https://www.moonshot.cn","models":[{"id":"kimi-k3","name":"Kimi K3","inputPrice":3,"outputPrice":15,"contextWindow":1048576,"released":"2026-07","openSource":true,"license":"Kimi K3 License","capabilities":["text","vision","code","tool-use","reasoning"],"tier":"flagship"}]},{"id":"fireworks","name":"Fireworks AI","logo":"/images/providers/fireworks.png","url":"https://fireworks.ai","models":[{"id":"ember-1","name":"Ember-1","inputPrice":3,"outputPrice":15,"contextWindow":1048576,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"}]},{"id":"sakana","name":"Sakana AI","logo":"/images/providers/sakana.png","url":"https://sakana.ai","models":[{"id":"fugu-ultra-v2","name":"Fugu Ultra v2.0","inputPrice":5,"outputPrice":30,"contextWindow":1000000,"released":"2026-09","capabilities":["text","vision","tool-use","code","reasoning"],"tier":"flagship"},{"id":"fugu-max","name":"Fugu Max","inputPrice":2,"outputPrice":6,"contextWindow":1000000,"released":"2026-09","capabilities":["text","tool-use","code","reasoning"],"tier":"mid"}]},{"id":"poolside","name":"poolside","logo":"/images/providers/poolside.png","url":"https://poolside.ai","models":[{"id":"laguna-s-2-1","name":"Laguna S 2.1","inputPrice":0.1,"outputPrice":0.2,"contextWindow":1048576,"released":"2026-07","openSource":true,"license":"OpenMDW-1.1","capabilities":["text","code","tool-use","reasoning"],"tier":"mid"}]},{"id":"unbiased","name":"Unbiased","logo":"/images/providers/unbiased.png","url":"https://unbiased.ai","models":[{"id":"pareto","name":"Pareto","inputPrice":2.5,"outputPrice":7.5,"contextWindow":262144,"released":"2026-09","capabilities":["text","vision","tool-use","code"],"tier":"flagship"}]},{"id":"tencent","name":"Tencent","logo":"/images/providers/tencent.png","url":"https://aistudio.tencent.com","models":[{"id":"hy4-preview","name":"Hy4 preview","inputPrice":0.834,"outputPrice":2.501,"contextWindow":1048576,"released":"2026-08","openSource":true,"license":"Apache-2.0","capabilities":["text","code","tool-use","reasoning"],"tier":"flagship"}]},{"id":"xiaomi","name":"Xiaomi","logo":"/images/providers/xiaomi.png","url":"https://mimo.xiaomi.com","models":[{"id":"mimo-v2-6-pro","name":"MiMo-V2.6-Pro","inputPrice":0.43,"outputPrice":0.87,"contextWindow":1048576,"released":"2026-09","openSource":true,"license":"MIT","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"flagship"},{"id":"mimo-v2-6-flash","name":"MiMo-V2.6-Flash","inputPrice":0.14,"outputPrice":0.28,"contextWindow":1048576,"released":"2026-09","openSource":true,"license":"MIT","capabilities":["text","vision","audio","video","tool-use","code","reasoning"],"tier":"budget"}]}],"pricingNotes":{"unit":"per 1M tokens","currency":"USD","openSourceNote":"Open source models are free to download and self-host. Hosted API pricing varies by provider.","disclaimer":"Prices are subject to change. Check provider websites for the most current pricing."}}}

2026-10-12/probe.jsonl ADDED Viewed

	@@ -0,0 +1 @@

+ {"ok":true,"summary":{"computed_at":"2026-10-12T17:15:56.163Z","window_label":"last_24h","providers":[{"provider":"anthropic","count":96,"success_count":96,"ok_pct":1,"measurable_count":96,"provider_ok_pct":1,"ttfb":{"p50":681,"p95":764,"p99":1747},"total":{"p50":681,"p95":764,"p99":1747},"status_codes":{"200":96},"last_probe_at":"2026-10-12T17:15:56.675Z","last_error":null,"probe_signal":{"signal":"healthy","window_minutes":60,"window_count":4,"provider_fails":0,"our_fails":0},"probe_recovery":{"all_ok":true,"window_minutes":20,"window_count":2,"last_ok_at":"2026-10-12T17:15:56.675Z"}},{"provider":"google","count":96,"success_count":26,"ok_pct":0.2708333333333333,"measurable_count":29,"provider_ok_pct":0.896551724137931,"ttfb":{"p50":474,"p95":986,"p99":1128},"total":{"p50":474,"p95":986,"p99":1128},"status_codes":{"0":3,"200":26,"429":67},"last_probe_at":"2026-10-12T17:15:56.675Z","last_error":"{\n \"error\": {\n \"code\": 429,\n \"message\": \"You exceeded your current quota, please check your plan and billing details. For more information on this error, head to: https://ai.google.dev/gemini-a","probe_signal":{"signal":"our_probe_limited","window_minutes":60,"window_count":4,"provider_fails":0,"our_fails":4},"probe_recovery":{"all_ok":false,"window_minutes":20,"window_count":2,"last_ok_at":null}},{"provider":"mistral","count":96,"success_count":0,"ok_pct":0,"measurable_count":0,"provider_ok_pct":null,"ttfb":{"p50":null,"p95":null,"p99":null},"total":{"p50":null,"p95":null,"p99":null},"status_codes":{"429":96},"last_probe_at":"2026-10-12T17:15:56.675Z","last_error":"{\"object\":\"error\",\"message\":\"Rate limit exceeded\",\"type\":\"rate_limited\",\"param\":null,\"code\":\"1300\",\"raw_status_code\":429}","probe_signal":{"signal":"our_probe_limited","window_minutes":60,"window_count":4,"provider_fails":0,"our_fails":4},"probe_recovery":{"all_ok":false,"window_minutes":20,"window_count":2,"last_ok_at":null}},{"provider":"cohere","count":96,"success_count":0,"ok_pct":0,"measurable_count":0,"provider_ok_pct":null,"ttfb":{"p50":null,"p95":null,"p99":null},"total":{"p50":null,"p95":null,"p99":null},"status_codes":{"429":96},"last_probe_at":"2026-10-12T17:15:56.675Z","last_error":"{\"id\":\"ed64ae36-90c9-42a4-9ff2-b1fde3166755\",\"message\":\"You are using a Trial key, which is limited to 1000 API calls / month. You can continue to use the Trial key for free or upgrade to a Production","probe_signal":{"signal":"our_probe_limited","window_minutes":60,"window_count":4,"provider_fails":0,"our_fails":4},"probe_recovery":{"all_ok":false,"window_minutes":20,"window_count":2,"last_ok_at":null}}]}}

2026-10-12/public-leaderboards.jsonl ADDED Viewed

	@@ -0,0 +1,26 @@

+{"id":"lmsys-arena","name":"Arena (formerly LMArena, originally LMSYS Chatbot Arena)","publisher":"Arena Intelligence","scope":"General chat capability across all major models, plus separate arenas for agents, web dev, vision, documents, search, image, and video","updateCadence":"continuous (live)","scoreType":"Elo-style rating (human pairwise vote)","domain":"general","live":true,"hasAPI":false,"url":"https://arena.ai/leaderboard/text","notes":"The most-cited crowd-sourced model ranking. The lmarena.ai domain now redirects to arena.ai. The Text arena alone passed 8.1M human pairwise votes by September 2026, with separate boards for Agent, WebDev, Vision, Document, Search, Text-to-Image, Image Edit, and video tasks."}
+{"id":"artificial-analysis","name":"Artificial Analysis","publisher":"Artificial Analysis","scope":"Quality + price + latency across models and providers","updateCadence":"daily","scoreType":"Composite quality index","domain":"general","live":true,"hasAPI":true,"url":"https://artificialanalysis.ai","notes":"Independent benchmark and market-data aggregator that runs evaluations itself. Its Intelligence Index (v4.3 as of September 2026) folds in agentic evals such as GDPval-AA v2.1, AutomationBench, and Terminal-Bench 4.0 alongside latency and price. Claude Opus 5.5 (max) took the top spot at 58 on 2026-09-22, ahead of GPT-6 Astra and Claude Fable 5.1 tied at 53; its Opus 5.5 runs used Anthropic's default fallback. Free data API with a key."}
+{"id":"hf-open-llm","name":"Open LLM Leaderboard","publisher":"Hugging Face","scope":"Open-weights models across 6 academic benchmarks (IFEval, BBH, MATH, GPQA, MUSR, MMLU-Pro)","updateCadence":"archived (no new evaluations)","scoreType":"Average of normalized benchmarks","domain":"open-models","live":false,"hasAPI":true,"url":"https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard","notes":"Archived. The Space still loads with an archived badge and the results datasets remain downloadable, but its last content update was March 2025 and later commits only fix startup crashes. Useful as a historical snapshot of 2024 era open models, not for current rankings."}
+{"id":"swebench-verified","name":"SWE-bench Verified Leaderboard","publisher":"SWE-bench team","scope":"Coding agents on 500 real GitHub issues","updateCadence":"as submissions land","scoreType":"% issues resolved","domain":"code","live":true,"hasAPI":false,"url":"https://www.swebench.com","notes":"The official board still tops out at 79.2% (Claude 4.5 Opus agent scaffolds, tied) and has taken no new Verified entry since February 2026; in September 2026 the separate bash-only board became a toggle inside Verified. Vendor launches and independent evaluators now report low to mid 90s on their own harnesses, so check the harness before comparing. Vals AI archived its independent Verified board as saturated on 2026-09-01, so new releases no longer get a Vals score either. TensorFeed mirrors top entries at /harnesses."}
+{"id":"aider-leaderboard","name":"Aider Polyglot Leaderboard","publisher":"Aider","scope":"225 hardest Exercism exercises across 6 languages","updateCadence":"dormant (page last updated November 2025)","scoreType":"% pass@2","domain":"code","live":false,"hasAPI":false,"url":"https://aider.chat/docs/leaderboards/","notes":"Edit-by-diff coding benchmark and a strong cross-language test, maintained by the Aider community. Frozen with gpt-5 (high) on top at 88.0%; no 2026 models have been added."}
+{"id":"livecodebench","name":"LiveCodeBench","publisher":"UC Berkeley + UW","scope":"Competitive programming problems released after model training cutoff","updateCadence":"no longer updating (last repo commit August 2025)","scoreType":"% pass@1","domain":"code","live":false,"hasAPI":false,"url":"https://livecodebench.github.io/leaderboard.html","notes":"Built to be contamination-free by pulling problems from after each model training cutoff, but the problem window stopped in spring 2025, so that guarantee no longer holds for current models. Vendors still quote it in launch tables; treat those figures as self-reported."}
+{"id":"bigcodebench","name":"BigCodeBench","publisher":"BigCode","scope":"1140 hard programming tasks with library usage","updateCadence":"not maintained (last results April 2025)","scoreType":"% pass@1","domain":"code","live":false,"hasAPI":false,"url":"https://huggingface.co/spaces/bigcode/bigcodebench-leaderboard","notes":"Multi-library code benchmark that tests realistic code composing external libraries; harder than HumanEval or MBPP for the same models. The board has not added results since April 2025."}
+{"id":"terminal-bench","name":"Terminal-Bench Leaderboard","publisher":"Terminal-Bench Team (Stanford, Laude Institute; Harbor harness)","scope":"Model plus agent-harness pairs on containerized terminal tasks with deterministic verifiers","updateCadence":"as runs land; versioned releases (4.0 current since August 2026)","scoreType":"% resolution rate with 95% interval","domain":"agent","live":true,"hasAPI":false,"url":"https://www.tbench.ai","notes":"Tests the loop, not just the model. The site now defaults to Terminal-Bench 4.0 (66 tasks) and also hosts 1.0, 2.0, 2.1, 3.0, and Science 0.1; versions are not comparable. On 4.0, Codex with GPT-6 Astra (max) leads at 58.2% as of September 2026, with Claude Code plus Claude Fable 5.1 at 57.9% inside the interval. Claude Opus 5.5 has no row on the official board yet; Vals AI's independent Terminus 2 run puts it first at 61.62%."}
+{"id":"frontiercode","name":"FrontierCode Leaderboard","publisher":"Cognition","scope":"Coding agents on maintainer-authored open-source tasks, graded for mergeability (Main 100 tasks, Extended 150)","updateCadence":"as models release (changelog on page)","scoreType":"Composite score (tests plus maintainer rubric), pass rate shown separately","domain":"code","live":true,"hasAPI":true,"url":"https://cognition.com/frontiercode","notes":"Grades whether a patch would actually be merged, not just whether tests pass; runs that consult solution-bearing sources are zeroed. Version 1.1 deprecated the Diamond subset. Per-effort results are published as a public JSON file behind the page. Claude Opus 5.5 (medium) took the top spot on 1.1 Main at 54.6 on 2026-09-22, about a point ahead of Claude Fable 5 (53.5), Claude Opus 5 (53.4), and GPT-6 Astra (53.3)."}
+{"id":"deepswe","name":"DeepSWE Leaderboard","publisher":"Datacurve","scope":"113 original long-horizon engineering tasks across 91 repos and 5 languages","updateCadence":"as models release","scoreType":"Pass@1 with error bars, plus cost, output tokens, and steps","domain":"code","live":true,"hasAPI":false,"url":"https://deepswe.datacurve.ai","notes":"Every model runs on mini-swe-agent for consistency, so it compares models rather than harnesses. Tasks are written from scratch to avoid contamination. A three-way tie at 74% (GPT-6 Astra, Gemini 3.8 Flash, Claude Opus 5) leads v1.1 as of September 2026."}
+{"id":"arc-prize","name":"ARC Prize Leaderboard","publisher":"ARC Prize Foundation","scope":"Abstract reasoning: ARC-AGI-1, ARC-AGI-2, and the interactive ARC-AGI-3 environments","updateCadence":"as verified runs land","scoreType":"% solved, plotted against cost per task","domain":"reasoning","live":true,"hasAPI":true,"url":"https://arcprize.org/leaderboard","notes":"Opens on ARC-AGI-3 and splits Verified from Community results; board data is published as JSON. As of September 2026, GPT-6 Astra (max) scores 95.0% on ARC-AGI-2 against a 100% human panel, and 62.7% on ARC-AGI-3 with the standard harness (99.9% with the OpenAI provider adapter harness). ARC-AGI-2 is now near saturation."}
+{"id":"mmlu-pro","name":"MMLU-Pro Leaderboard","publisher":"TIGER Lab","scope":"General knowledge + reasoning across 57 subjects","updateCadence":"occasional (results dataset last updated March 2026)","scoreType":"% accuracy","domain":"general","live":true,"hasAPI":false,"url":"https://huggingface.co/spaces/TIGER-Lab/MMLU-Pro","notes":"Successor to MMLU and the standard knowledge plus reasoning check of 2024 to 2026. The maintainer board updates slowly, and Vals AI archived its independent MMLU Pro board as saturated on 2026-09-01, so models released after that date may have no independent score at all."}
+{"id":"hle-leaderboard","name":"Humanity's Last Exam","publisher":"CAIS + Scale AI","scope":"2,500 expert-validated questions across 100+ disciplines (finalized April 2025)","updateCadence":"lastexam.ai table is static; the live board is at Scale Labs","scoreType":"% accuracy","domain":"reasoning","live":true,"hasAPI":false,"url":"https://labs.scale.com/leaderboard/humanitys_last_exam","notes":"Every major lab reports against HLE, with and without tools. The original lastexam.ai page still shows a 2025 table; the maintained ranking lives on Scale Labs, where GPT-6 Astra leads at 54.8% and the next entries sit in the mid 40s as of September 2026. Tool setups vary by vendor, so compare like for like."}
+{"id":"mmmu","name":"MMMU Leaderboard","publisher":"TIGER Lab","scope":"College-level multimodal questions across 30 subjects","updateCadence":"as submissions land (last entry July 2026)","scoreType":"% accuracy","domain":"multimodal","live":true,"hasAPI":false,"url":"https://mmmu-benchmark.github.io","notes":"Standard vision-language benchmark with MMMU (val) and MMMU-Pro tables. The board aggregates submitted results rather than re-running models. Top MMMU-Pro entry is 86.9 and top MMMU val is 85.4 as of September 2026."}
+{"id":"video-arena","name":"Artificial Analysis Video Arena","publisher":"Artificial Analysis","scope":"Video generation models, ranked separately with and without audio output","updateCadence":"continuous","scoreType":"Elo (human pairwise vote)","domain":"video","live":true,"hasAPI":false,"url":"https://artificialanalysis.ai/video/leaderboard/text-to-video","notes":"Live Elo for video models. The old /text-to-video/arena path now opens the voting page, so link the leaderboard path instead. Wan 3.0 leads the with-audio text-to-video board at 1242 as of September 2026; HappyHorse 1.0 no longer leads."}
+{"id":"image-arena","name":"Artificial Analysis Image Arena","publisher":"Artificial Analysis","scope":"Image generation and image editing models","updateCadence":"continuous","scoreType":"Elo (human pairwise vote)","domain":"image","live":true,"hasAPI":false,"url":"https://artificialanalysis.ai/image/leaderboard/text-to-image","notes":"Live Elo for image models. The old /text-to-image/arena path now opens the voting page. GPT Image 2.5 Flare (max) leads text-to-image at 1187 as of September 2026."}
+{"id":"tts-arena","name":"TTS Arena V2","publisher":"TTS-AGI (Hugging Face Space)","scope":"TTS models from commercial and open providers, including stealth entries","updateCadence":"continuous","scoreType":"Elo (human pairwise vote)","domain":"voice","live":true,"hasAPI":true,"url":"https://huggingface.co/spaces/TTS-AGI/TTS-Arena-V2","notes":"Live Elo for TTS quality with a public JSON leaderboard endpoint. The original TTS Arena Space is now a read-only legacy board. A stealth model leads V2 as of September 2026 and Eleven v3 sits around 27th. TensorFeed mirrors top entries at /voice-leaderboards."}
+{"id":"open-asr","name":"Open ASR Leaderboard","publisher":"Hugging Face","scope":"STT models on LibriSpeech + Common Voice + AMI + GigaSpeech","updateCadence":"as submissions land","scoreType":"Word Error Rate","domain":"voice","live":true,"hasAPI":true,"url":"https://huggingface.co/spaces/hf-audio/open_asr_leaderboard","notes":"Aggregated ASR benchmark with English and multilingual tracks, actively updated through September 2026. The English average WER leader has turned over repeatedly since 2025, and AssemblyAI Universal-2 no longer leads."}
+{"id":"ruler-leaderboard","name":"RULER Leaderboard","publisher":"NVIDIA","scope":"Long-context retrieval and reasoning (effective vs claimed length)","updateCadence":"README table frozen since October 2025 (repo still receives fixes)","scoreType":"% accuracy at varying context lengths","domain":"long-context","live":false,"hasAPI":false,"url":"https://github.com/NVIDIA/RULER","notes":"Reveals the gap between claimed and effective context length. The README leaderboard has not been updated since RULER v2 was added in October 2025, so it predates the 1M-token context flagships of 2026."}
+{"id":"gaia-leaderboard","name":"GAIA Leaderboard","publisher":"Hugging Face + Meta","scope":"General assistant agents on 466 real-world questions","updateCadence":"as submissions land","scoreType":"% accuracy by difficulty level","domain":"agent","live":true,"hasAPI":false,"url":"https://huggingface.co/spaces/gaia-benchmark/leaderboard","notes":"Tests the full agent loop: web browsing, file ops, multi-step reasoning, at three difficulty levels. Largely saturated: the test-split leader, an agent scaffold, scores 93.36% as of September 2026."}
+{"id":"webarena","name":"WebArena Leaderboard","publisher":"CMU + UW","scope":"Browser agents on 812 tasks across 5 simulated websites","updateCadence":"as submissions land (maintained as a linked spreadsheet)","scoreType":"% solved","domain":"agent","live":true,"hasAPI":false,"url":"https://webarena.dev/og/","notes":"The browser-automation agent benchmark. The old /leaderboard path now returns 404 and webarena.dev is a WebArena-x hub; the original board is a spreadsheet linked from the /og/ page. Frontier is 74.3% (WebTactix with DeepSeek v3.2) as of September 2026."}
+{"id":"osworld-leaderboard","name":"OSWorld-Verified Leaderboard (OSWorld 1.0)","publisher":"XLANG Lab, University of Hong Kong","scope":"Computer-use agents on 369 desktop tasks (361 excluding Google Drive tasks)","updateCadence":"as verified submissions land (newest entry August 2026)","scoreType":"% solved","domain":"agent","live":true,"hasAPI":false,"url":"https://osworld-v1.xlang.ai","notes":"The original computer-use benchmark, now near saturation: the Intelligence-Indeed Agent scaffold leads OSWorld-Verified at 90.19% and the best bare general model sits in the mid 80s. os-world.github.io redirects here. For long-horizon computer use, see the separate OSWorld 2.0 board."}
+{"id":"osworld-2","name":"OSWorld 2.0 Leaderboard","publisher":"XLANG Lab, University of Hong Kong","scope":"Computer-use agents on 108 long-horizon workflows across 7 professional domains","updateCadence":"as official runs land","scoreType":"Binary completion and partial score at a fixed step budget","domain":"agent","live":true,"hasAPI":true,"url":"https://osworld-v2.xlang.ai","notes":"Filter by task release (v2026.06.24 or v2026.08.08), full set or offline no-internet subset, and step budget before comparing anything; vendor launch figures often mix these. Official results are served as a public JSON file. Claude Opus 5 (max) leads the v2026.08.08 full set at 31.43% binary as of September 2026."}
+{"id":"automationbench","name":"AutomationBench Leaderboard","publisher":"Zapier","scope":"Agents completing cross-app business workflows across 47 simulated SaaS tools in 6 domains","updateCadence":"as models release (versioned, currently 1.0.6)","scoreType":"% tasks completed correctly (strict end-state assertions)","domain":"agent","live":true,"hasAPI":false,"url":"https://zapier.com/benchmarks","notes":"Scored on a held-out private task set with deterministic checks and no LLM judge; the public GitHub task set is easier. Lists cost per task and per-domain leaders. Watch for fallback rows, such as Claude Fable 5.1 with Claude Opus 5 completing refused steps. Claude Opus 5.5 (max, default fallbacks) took the lead at 42.47% on 2026-09-22, just ahead of GPT-6 Astra (max) at 41.4%."}
+{"id":"steel-browsecomp","name":"BrowseComp Leaderboard","publisher":"Steel","scope":"Agentic web research systems on OpenAI's BrowseComp","updateCadence":"as results are published","scoreType":"% accuracy","domain":"agent","live":true,"hasAPI":false,"url":"https://leaderboard.steel.dev/leaderboards/browsecomp/","notes":"BrowseComp has no official leaderboard, so this board collects self-reported figures with a source link and setup notes per row (single agent vs multi-agent, context management). Treat it as a registry of claims, not a controlled comparison."}
+{"id":"vals-ai","name":"Vals AI Benchmarks","publisher":"Vals AI","scope":"Independent runs of frontier models across academic, legal, finance, healthcare, and coding benchmarks, plus the composite Vals Index","updateCadence":"weekly to monthly per benchmark (dated on each board)","scoreType":"Accuracy per benchmark; Vals Index composite","domain":"general","live":true,"hasAPI":false,"url":"https://www.vals.ai/benchmarks","notes":"Runs every model itself on a fixed harness, so its numbers often differ from vendor launch claims. Each board shows its own last-updated date and model count. Claude Opus 5.5 leads the Vals Index at 69.69% as of September 2026, ahead of Claude Fable 5.1 (68.83%). On 2026-09-01 Vals archived SWE-bench Verified (run bash-only on mini-swe-agent), MMLU Pro, and GPQA Diamond as saturated and no longer runs them on new releases."}

2026-10-12/specialized-models.jsonl ADDED Viewed

	@@ -0,0 +1,24 @@

+{"id":"codestral-25.08","name":"Codestral 25.08","publisher":"Mistral","domain":"code","params":"undisclosed","pricing":"$0.30 input / $0.90 output per 1M tokens via Mistral API","openWeights":false,"license":"Mistral commercial (Premier model)","released":"2025-07","benchmark":null,"capabilities":["code completion","fill-in-the-middle","chat + function calling","128k context"],"url":"https://docs.mistral.ai/models/codestral-25-08","notes":"Mistral's low-latency completion model and the current Codestral; Codestral 25.01 was retired from the API on 2025-11-30. Mistral reports +30% accepted completions and 50% fewer runaway generations over prior versions but publishes no absolute benchmark scores. Deployable in cloud, VPC, or on-prem under enterprise terms. For agentic coding, Mistral now routes former Devstral users to Mistral Medium 3.5."}
+{"id":"mai-code-1.1-flash","name":"MAI-Code-1.1-Flash","publisher":"Microsoft AI","domain":"code","params":"138B MoE / 5B active","pricing":"$0.20 input / $1.20 output per 1M tokens (GitHub Copilot list rate)","openWeights":false,"license":"Proprietary (GitHub Copilot terms)","released":"2026-08","benchmark":"SWE-Bench Verified 72.6%, Terminal-Bench 2.1 62.9% (Copilot harness, vendor-run)","capabilities":["agentic coding","image input","256k context","GitHub Copilot integration"],"url":"https://github.com/microsoft/MAI-Code","notes":"Microsoft AI's in-house small-tier coding model, trained against the GitHub Copilot production harness and distributed only inside Copilot clients. Succeeds MAI-Code-1-Flash (June 2026) at a quarter of its list price and adds vision input. All published scores come from Microsoft's own Copilot harness, so treat cross-lab comparisons with care."}
+{"id":"qwen-coder-2.5-32b","name":"Qwen 2.5 Coder 32B","publisher":"Alibaba","domain":"code","params":"32B","pricing":"Open weights, free to self-host","openWeights":true,"license":"Apache-2.0","released":"2024-11","benchmark":"HumanEval 92.7%, MBPP 90.2%","capabilities":["code completion","code repair","40+ languages","128k context"],"url":"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct","notes":"Apache-2.0 dense code model that punches above its parameter count. Still a solid self-hosted default, though the Qwen line has moved on to Qwen3-Coder (480B-A35B and 30B-A3B) and Qwen3-Coder-Next (80B MoE / 3B active, Apache-2.0, Feb 2026) for agentic coding."}
+{"id":"starcoder-2-15b","name":"StarCoder 2 15B","publisher":"BigCode","domain":"code","params":"15B","pricing":"Open weights, free to self-host","openWeights":true,"license":"BigCode OpenRAIL-M","released":"2024-02","benchmark":"HumanEval 46.3% pass@1","capabilities":["code completion","fill-in-the-middle","600+ languages","16k context"],"url":"https://huggingface.co/bigcode/starcoder2-15b","notes":"Reproducible: trained on The Stack v2 (open dataset). Best fit for academic / regulated code-completion deployments."}
+{"id":"med-gemini","name":"Med-Gemini","publisher":"Google","domain":"medical","params":"undisclosed","pricing":"Not publicly available (research model)","openWeights":false,"license":"Proprietary","released":"2024-04","benchmark":"MedQA-USMLE 91.1%","capabilities":["clinical Q&A","medical imaging","EHR summarization","long-context records"],"url":"https://research.google/blog/advancing-medical-ai-with-med-gemini/","notes":"Google's research family of Gemini models fine-tuned for medicine (radiology, pathology, dermatology, genomics). Set a state-of-the-art 91.1% on MedQA at publication. Google says considerable further research is needed before real-world clinical use, so it is not a self-serve product; Google's downloadable medical line is MedGemma."}
+{"id":"apertus-70b-meditronfo","name":"Apertus-70B-MeditronFO","publisher":"EPFL (LiGHT lab)","domain":"medical","params":"70B","pricing":"Open weights, free to self-host","openWeights":true,"license":"Apache-2.0","released":"2026-05","benchmark":"MedQA 68.6%, HealthBench Hard 40.1%","capabilities":["clinical Q&A","guideline-grounded answers","differential diagnosis"],"url":"https://huggingface.co/EPFLiGHT/Apertus-70B-MeditronFO","notes":"Flagship of EPFL's Fully Open Meditron family: open weights, open training data, and an open recipe, fine-tuned from Swiss AI's Apertus-70B-Instruct on a clinician-vetted corpus built partly from 46,469 clinical practice guidelines. Siblings ship on Apertus 8B, Gemma 3 27B, EuroLLM, and OLMo 2 bases. Succeeds the Llama 2 based Meditron 70B as EPFL's reference open medical model."}
+{"id":"biomistral","name":"BioMistral 7B","publisher":"BioMistral collaboration","domain":"medical","params":"7B","pricing":"Open weights, free to self-host","openWeights":true,"license":"Apache-2.0","released":"2024-02","benchmark":"MedQA-USMLE 50.6% (4-option)","capabilities":["biomedical Q&A","drug interactions","literature retrieval"],"url":"https://huggingface.co/BioMistral/BioMistral-7B","notes":"Mistral 7B continued-pretrained on PubMed. Apache-licensed; small enough to self-host on a single consumer GPU. Good base for medical RAG agents."}
+{"id":"thomson-1.0-small","name":"Thomson-1.0-Small","publisher":"Thomson Reuters","domain":"legal","params":"35B MoE / 3B active","pricing":"Open weights, free for non-commercial use","openWeights":true,"license":"PolyForm Strict 1.0.0 (non-commercial)","released":"2026-08","benchmark":"Harvey Legal Agent Benchmark 73.4%, LegalBench 79.9% (vendor harness)","capabilities":["legal research agents","contract understanding","tax Q&A","262k context"],"url":"https://huggingface.co/thomsonreuters/Thomson-1.0-Small","notes":"Open-weight member of Thomson Reuters' Thomson-1.0 family, continually trained from Qwen3.6-35B-A3B on proprietary case law, statutes, contracts, filings, and news. The larger proprietary Thomson model powers Tabular Analysis in CoCounsel Legal. Gains show up on agentic and document tasks; on plain LegalBench it lands slightly below its Qwen base."}
+{"id":"saul-lm-141b","name":"SaulLM 141B","publisher":"Equall","domain":"legal","params":"141B (Mixtral-based)","pricing":"Open weights, free to self-host","openWeights":true,"license":"MIT","released":"2024-07","benchmark":"Beat prior open models on LegalBench-Instruct at release (paper)","capabilities":["legal Q&A","contract review","case law summarization","multilingual legal"],"url":"https://huggingface.co/Equall/SaulLM-141B-Instruct","notes":"Largest open legal LLM at its release. Continued pretraining of Mixtral on a 540B+ token legal corpus, released in base, instruct, and aligned versions. Trained on US and European legal texts; a research artifact, not legal advice."}
+{"id":"saul-lm-7b","name":"SaulLM 7B","publisher":"Equall","domain":"legal","params":"7B","pricing":"Open weights, free to self-host","openWeights":true,"license":"MIT","released":"2024-03","benchmark":null,"capabilities":["legal Q&A","contract clauses","case briefing"],"url":"https://huggingface.co/Equall/Saul-7B-Base","notes":"Smaller SaulLM, continued pretraining of Mistral 7B on 30B+ tokens of English legal text. Runs on consumer GPU at Q4. Solid base for legal RAG agents."}
+{"id":"fingpt-v3","name":"FinGPT v3","publisher":"AI4Finance Foundation","domain":"finance","params":"6B-13B (LoRA on ChatGLM2 / Llama 2)","pricing":"Open weights, free to self-host","openWeights":true,"license":"MIT","released":"2023-10","benchmark":"FPB weighted F1 0.882 (v3.3)","capabilities":["financial sentiment","news + tweet classification","single-GPU LoRA fine-tuning"],"url":"https://github.com/AI4Finance-Foundation/FinGPT","notes":"Open finance models maintained by AI4Finance Foundation. v3 is a set of LoRA adapters: v3.1 on ChatGLM2-6B, v3.2 on Llama 2 7B, v3.3 on Llama 2 13B, trainable on a single RTX 3090. Best fit for sentiment + classification rather than reasoning."}
+{"id":"bloomberggpt","name":"BloombergGPT","publisher":"Bloomberg","domain":"finance","params":"50B","pricing":"Internal Bloomberg use only (not public)","openWeights":false,"license":"Proprietary","released":"2023-03","benchmark":"FPB weighted F1 51.07 (5-shot)","capabilities":["financial NER","sentiment","headline classification","financial QA (ConvFinQA)"],"url":"https://www.bloomberg.com/company/press/bloomberggpt-50-billion-parameter-llm-tuned-finance/","notes":"Bloomberg's in-house finance LLM. Trained on a private 363B-token finance corpus plus public data. Reference for \"what does a finance-specialized base model look like\" but not externally usable."}
+{"id":"suno-v6","name":"Suno v6","publisher":"Suno","domain":"music","params":"undisclosed","pricing":"Free tier (v6-mini); Pro $8/mo, Premier $24/mo billed yearly","openWeights":false,"license":"Proprietary","released":"2026-09","benchmark":null,"capabilities":["lyrics + music","genre control","v6-wild exploratory variant","free v6-mini tier"],"url":"https://suno.com/blog/introducing-v6","notes":"Suno's first model generation developed with industry partners including Warner Music Group, BMG, and Believe. Ships as v6 (flagship, paid), v6-wild (less predictable, paid), and v6-mini (free). Suno is retiring its earlier models as v6 rolls out. The default for music-gen agents that need polished output."}
+{"id":"udio","name":"Udio","publisher":"Uncharted Labs","domain":"music","params":"undisclosed","pricing":"Subscription plans (see udio.com/pricing)","openWeights":false,"license":"Proprietary","released":"2024-04","benchmark":null,"capabilities":["music gen","extend","remix","lyrics or instrumental"],"url":"https://www.udio.com","notes":"Suno competitor that settled with Universal Music Group (Oct 2025) and Warner Music Group (Nov 2025). Since the UMG deal, creations stay inside a walled garden and cannot be downloaded. A licensed subscription platform trained on authorized music is in development, so it is a poor fit for agents that need exportable audio today."}
+{"id":"minimax-music-3","name":"MiniMax Music 3","publisher":"MiniMax","domain":"music","params":"8B global LM + 0.6B local LM + 2.4B flow-matching decoder","pricing":"Open weights, free to self-host","openWeights":true,"license":"MiniMax-Music3 Community License","released":"2026-08","benchmark":null,"capabilities":["lyrics + music","full songs up to 5 min","structured style captions","32 kHz stereo output"],"url":"https://huggingface.co/MiniMaxAI/MiniMax-Music3","notes":"Open full-song model with vocals: an 8B global LLM (initialized from Qwen3-8B) plans structure, a 0.6B local LLM fills acoustic detail, and a flow-matching decoder renders audio. Fits under 24GB VRAM, or 8GB with offloading. Commercial use requires on-screen attribution, and products above $20M yearly revenue need written authorization."}
+{"id":"musicgen-large","name":"MusicGen Large","publisher":"Meta","domain":"music","params":"3.3B","pricing":"Open weights, free to self-host","openWeights":true,"license":"CC-BY-NC-4.0","released":"2023-06","benchmark":null,"capabilities":["instrumental music gen","melody conditioning","audio continuation"],"url":"https://huggingface.co/facebook/musicgen-large","notes":"Open instrumental music model trained on licensed Meta, Shutterstock, and Pond5 music. Non-commercial weights license. Now an older research baseline; newer open options include Stable Audio 3.0 and MiniMax Music 3."}
+{"id":"stable-audio-3","name":"Stable Audio 3.0","publisher":"Stability AI","domain":"music","params":"1.4B (Medium) / 2.7B (Large) diffusion transformer","pricing":"Small + Medium open weights; Large via API at 26 credits ($0.26) per generation","openWeights":true,"license":"Stability AI Community License (Enterprise License above $1M revenue)","released":"2026-05","benchmark":null,"capabilities":["music + SFX gen","up to 6m20s clips","audio-to-audio","inpainting / continuation"],"url":"https://stability.ai/news-updates/meet-stable-audio-3-the-model-family-built-for-artistic-experimentation-with-open-weight-models","notes":"Stability's audio family trained on licensed AudioSparx audio plus filtered Creative Commons Freesound recordings. Small variants run on CPU and phones, Medium needs a CUDA GPU, Large is API and enterprise only. Designed for sound design workflows; replaces Stable Audio 2.0, which is no longer on Stability's API price list."}
+{"id":"trellis","name":"TRELLIS","publisher":"Microsoft Research","domain":"3d","params":"up to 2B","pricing":"Open weights, free to self-host","openWeights":true,"license":"MIT","released":"2024-12","benchmark":null,"capabilities":["image to 3D","text to 3D","mesh + gaussian splat output","local 3D editing"],"url":"https://github.com/microsoft/TRELLIS","notes":"Microsoft Research 3D generative model built on a unified structured latent. Outputs textured mesh, Gaussian splats, or radiance fields. For image-to-3D, Microsoft's TRELLIS.2 (4B, MIT, Dec 2025) is the successor and adds full PBR materials."}
+{"id":"hunyuan3d-2","name":"Hunyuan3D-2","publisher":"Tencent","domain":"3d","params":"1.1B shape + 1.3B texture","pricing":"Open weights, free to self-host","openWeights":true,"license":"Tencent Hunyuan Community License","released":"2025-01","benchmark":null,"capabilities":["image to 3D mesh","texture generation","text to 3D"],"url":"https://huggingface.co/tencent/Hunyuan3D-2","notes":"Tencent's open 3D model. Two-stage: shape DiT + texture paint model. Strongest open 3D output quality in early 2025. Hunyuan3D-2.1 (June 2025) followed with PBR texture synthesis and released training code."}
+{"id":"colpali","name":"ColPali","publisher":"ILLUIN Technology","domain":"retrieval","params":"3B (PaliGemma-based)","pricing":"Open weights, free to self-host","openWeights":true,"license":"MIT","released":"2024-07","benchmark":"ViDoRe avg nDCG@5 81.3 (paper)","capabilities":["vision-document retrieval","late interaction","PDF / chart understanding"],"url":"https://huggingface.co/vidore/colpali","notes":"Vision-document retrieval model. Embeds whole-page images directly, skipping OCR. Strong on PDFs with tables, charts, scientific figures. Newer checkpoints (colpali-v1.3, ColQwen2.5) live under the same vidore org."}
+{"id":"evie-8b","name":"EVIE-8B","publisher":"Tencent","domain":"retrieval","params":"8.4B (Qwen3.5-based)","pricing":"Open weights, free to self-host","openWeights":true,"license":"Apache-2.0","released":"2026-09","benchmark":"ViDoRe V3 nDCG@10 66.75, ViDoRe V1 nDCG@5 92.18 (vendor-reported)","capabilities":["vision-document retrieval","late interaction","4096-dim token vectors","layout + table matching"],"url":"https://huggingface.co/tencent/EVIE-8B","notes":"ColPali-style multi-vector retriever on a Qwen3.5 backbone with bidirectional attention, which Tencent reports as rank 1 on ViDoRe V3 at release. Also serves as the teacher for EVIE-4.5B, which supports 64 to 2048 dim prefix embeddings for cheaper indexes. Paper still pending, so scores are self-reported."}
+{"id":"splade-v3","name":"SPLADE v3","publisher":"Naver Labs","domain":"retrieval","params":"110M","pricing":"Open weights, free to self-host","openWeights":true,"license":"CC-BY-NC-SA-4.0","released":"2024-03","benchmark":"BEIR-13 avg nDCG@10 51.7","capabilities":["sparse retrieval","BM25-compatible inverted index","lexical match expansion"],"url":"https://huggingface.co/naver/splade-v3","notes":"Learned sparse retrieval. Drops into existing inverted-index infra (Elasticsearch, Lucene) without dense vector overhead. Strong hybrid-search complement to dense embeddings. Gated on Hugging Face and non-commercial licensed."}
+{"id":"esmfold2","name":"ESMFold2","publisher":"Biohub","domain":"science","params":"Folding model on the 6B-parameter ESMC protein language model","pricing":"Open weights, free to self-host; also on Biohub Platform","openWeights":true,"license":"MIT","released":"2026-05","benchmark":"FoldBench: 55% antibody-antigen, 71% protein-protein (77% with MSA)","capabilities":["protein structure prediction","protein + DNA/RNA + ligand complexes","binder design","optional MSA input"],"url":"https://biohub.ai/esm/protein","notes":"Next-generation ESM structure model released by Biohub alongside ESMC and the ESM Atlas. Biohub reports it beats AlphaFold 3 at antibody-antigen pose prediction from single sequences, and used it to design lab-validated binders against five cancer and immunology targets. ESMFold2-Fast folds a 1,024-residue protein in 9.4 seconds."}
+{"id":"leanstral-1.5","name":"Leanstral 1.5","publisher":"Mistral","domain":"science","params":"119B MoE / 6.5B active","pricing":"Open weights; free API endpoint (labs-leanstral-1-5)","openWeights":true,"license":"Apache-2.0","released":"2026-07","benchmark":"miniF2F 100%, PutnamBench 587/672, FATE-H 87%","capabilities":["Lean 4 theorem proving","autoformalization","code verification agent","256k context"],"url":"https://mistral.ai/news/leanstral-1-5/","notes":"Mistral's formal-math specialist, built on the Mistral Small 4 family. Works as an agent inside a Lean project, reading compiler feedback and editing files across millions of tokens; Mistral puts PutnamBench cost at about $4 per problem. Mistral says it surfaced 5 previously unknown bugs across 57 repositories."}

2026-10-12/status.jsonl ADDED Viewed

	@@ -0,0 +1,23 @@

+{"name":"Claude API","provider":"Anthropic","status":"operational","statusPageUrl":"https://status.anthropic.com","components":[{"name":"claude.ai","status":"operational"},{"name":"Claude API (api.anthropic.com)","status":"operational"},{"name":"Claude Console (platform.claude.com)","status":"degraded"},{"name":"Claude Code","status":"operational"},{"name":"Claude Cowork","status":"operational"},{"name":"Claude for Government","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.633Z"}
+{"name":"OpenAI API","provider":"OpenAI","status":"operational","statusPageUrl":"https://status.openai.com","components":[{"name":"Chat Completions","status":"operational"},{"name":"Responses","status":"operational"},{"name":"Agents","status":"operational"},{"name":"Decisions","status":"operational"},{"name":"Fine-tuning","status":"operational"},{"name":"Embeddings","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.225Z"}
+{"name":"Google Gemini","provider":"Google","status":"operational","statusPageUrl":"https://status.cloud.google.com","components":[{"name":"Inference API","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.128Z"}
+{"name":"GitHub Copilot","provider":"GitHub","status":"operational","statusPageUrl":"https://www.githubstatus.com","components":[{"name":"Copilot","status":"operational"},{"name":"Copilot AI Model Providers","status":"operational"},{"name":"Git Operations","status":"operational"},{"name":"Webhooks","status":"operational"},{"name":"API Requests","status":"operational"},{"name":"Issues","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.132Z"}
+{"name":"Cursor","provider":"Cursor","status":"operational","statusPageUrl":"https://status.cursor.com","components":[{"name":"Automations","status":"operational"},{"name":"Review Agents","status":"operational"},{"name":"Cloud Agents","status":"operational"},{"name":"cursor.com","status":"operational"},{"name":"IDE","status":"operational"},{"name":"Origin","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.158Z"}
+{"name":"Perplexity","provider":"Perplexity AI","status":"operational","statusPageUrl":"https://status.perplexity.com","components":[{"name":"Website","status":"operational"},{"name":"Computer","status":"operational"},{"name":"API","status":"operational"},{"name":"App","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.332Z"}
+{"name":"Groq","provider":"Groq","status":"operational","statusPageUrl":"https://groqstatus.com","components":[{"name":"API","status":"operational"},{"name":"llama-3.1-8b-instant","status":"operational"},{"name":"llama-3.3-70b-versatile","status":"operational"},{"name":"whisper-large-v3","status":"operational"},{"name":"whisper-large-v3-turbo","status":"operational"},{"name":"Canopy Labs Orpheus Arabic Saudi","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.266Z"}
+{"name":"AWS Bedrock","provider":"AWS","status":"operational","statusPageUrl":"https://health.aws.amazon.com/health/status","components":[{"name":"Bedrock API","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.204Z"}
+{"name":"Azure OpenAI","provider":"Microsoft Azure","status":"operational","statusPageUrl":"https://azure.status.microsoft/en-us/status","components":[{"name":"Azure OpenAI Service","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.192Z"}
+{"name":"Hugging Face","provider":"Hugging Face","status":"operational","statusPageUrl":"https://status.huggingface.co","components":[{"name":"API","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.454Z"}
+{"name":"Replicate","provider":"Replicate","status":"operational","statusPageUrl":"https://status.replicate.com","components":[{"name":"Replicate","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.224Z"}
+{"name":"Cohere","provider":"Cohere","status":"operational","statusPageUrl":"https://status.cohere.com","components":[{"name":"All Endpoints","status":"operational"},{"name":"embeddings","status":"operational"},{"name":"All Models","status":"operational"},{"name":"command-a-plus-05-2026","status":"operational"},{"name":"command-a-reasoning-08-2025","status":"operational"},{"name":"command-a-vision-07-2025","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.375Z"}
+{"name":"Mistral","provider":"Mistral AI","status":"operational","statusPageUrl":"https://status.mistral.ai","components":[{"name":"API","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.326Z"}
+{"name":"DeepSeek","provider":"DeepSeek","status":"operational","statusPageUrl":"https://status.deepseek.com","components":[{"name":"API","status":"operational"}],"lastChecked":"2026-10-12T17:14:57.513Z"}
+{"name":"Together AI","provider":"Together AI","status":"operational","statusPageUrl":"https://status.together.ai","components":[{"name":"API","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.593Z"}
+{"name":"Fireworks AI","provider":"Fireworks AI","status":"operational","statusPageUrl":"https://status.fireworks.ai","components":[{"name":"API","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.914Z"}
+{"name":"OpenRouter","provider":"OpenRouter","status":"operational","statusPageUrl":"https://status.openrouter.ai","components":[{"name":"Chat (/api/v1/chat/completions)","status":"operational"},{"name":"Video (/api/v1/videos)","status":"operational"},{"name":"Image (/api/v1/image)","status":"operational"},{"name":"TTS (/api/v1/audio/speech)","status":"operational"},{"name":"STT (/api/v1/audio/transcriptions)","status":"operational"},{"name":"Embeddings (/api/v1/embeddings)","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.399Z"}
+{"name":"ElevenLabs","provider":"ElevenLabs","status":"operational","statusPageUrl":"https://status.elevenlabs.io","components":[{"name":"Text to Speech","status":"operational"},{"name":"Speech to Text","status":"operational"},{"name":"Other","status":"operational"},{"name":"Telephony","status":"operational"},{"name":"RAG","status":"operational"},{"name":"Quality","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.609Z"}
+{"name":"Stability AI","provider":"Stability AI","status":"operational","statusPageUrl":"https://status.stability.ai","components":[{"name":"Stability.ai Platform & Services","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.599Z"}
+{"name":"Runway","provider":"Runway","status":"operational","statusPageUrl":"https://status.runwayml.com","components":[{"name":"Backend","status":"operational"},{"name":"Support","status":"operational"},{"name":"Public API","status":"operational"},{"name":"App","status":"operational"},{"name":"Billing","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.455Z"}
+{"name":"Luma","provider":"Luma AI","status":"operational","statusPageUrl":"https://status.lumalabs.ai","components":[{"name":"API","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.618Z"}
+{"name":"Midjourney","provider":"Midjourney","status":"operational","statusPageUrl":"https://status.midjourney.com","components":[{"name":"Web","status":"operational"},{"name":"Bot","status":"operational"},{"name":"API","status":"operational"},{"name":"Image Generation - Fast & Turbo","status":"operational"},{"name":"Image Generation - Relax","status":"operational"},{"name":"Billing","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.835Z"}
+{"name":"Hyperliquid","provider":"Hyperliquid","status":"operational","statusPageUrl":"https://hyperliquid.statuspage.io","components":[{"name":"Frontend","status":"operational"},{"name":"Hyperliquid L1","status":"operational"},{"name":"API","status":"operational"}],"lastChecked":"2026-10-12T17:14:56.873Z"}

2026-10-12/training-datasets.jsonl ADDED Viewed

	@@ -0,0 +1,29 @@

+{"id":"fineweb","name":"FineWeb","publisher":"Hugging Face","stage":"pretraining","contentType":"web text","tokens":"18.5T","items":"25.9B documents","license":"ODC-BY-1.0","languages":"English","released":"2024-04","url":"https://huggingface.co/datasets/HuggingFaceFW/fineweb","notes":"Filtered and deduplicated CommonCrawl, processed with the datatrove library. Launched at 15T gpt2 tokens; v1.4.0 (July 2025) added the January to June 2025 snapshots and the card now lists more than 18.5T tokens. Some domains were removed in January 2025 in response to a cease-and-desist notice."}
+{"id":"fineweb-edu","name":"FineWeb-Edu","publisher":"Hugging Face","stage":"pretraining","contentType":"educational web text","tokens":"1.3T","items":"1.5B documents","license":"ODC-BY-1.0","languages":"English","released":"2024-05","url":"https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu","notes":"FineWeb subset kept by an educational-quality classifier trained on Llama-3-70B-Instruct annotations. A looser threshold variant (FineWeb-Edu-score-2) holds 5.4T tokens. Hugging Face ablations report it outperforming full FineWeb on knowledge and reasoning benchmarks such as MMLU and ARC."}
+{"id":"common-crawl","name":"Common Crawl","publisher":"Common Crawl Foundation","stage":"pretraining","contentType":"web text","tokens":"petabyte-scale raw","items":"300B+ webpages","license":"Common Crawl Terms of Use (limited license; crawled content keeps third-party rights)","languages":"multilingual (200+)","released":"ongoing, 3-5B new pages added monthly","url":"https://commoncrawl.org","notes":"The raw web crawl most open web pretraining datasets are built on. Not a public-domain dedication: the terms grant a limited license to the service and require users to respect the copyrights of the crawled material. Most labs filter it heavily before training."}
+{"id":"redpajama-v2","name":"RedPajama v2","publisher":"Together AI","stage":"pretraining","contentType":"web text","tokens":"30T","items":"100B+ documents","license":"Common Crawl Terms of Use (data); Apache-2.0 (code)","languages":"English, German, French, Spanish, Italian","released":"2023-10","url":"https://huggingface.co/datasets/togethercomputer/RedPajama-Data-V2","notes":"Built from 84 CommonCrawl snapshots with the CCNet pipeline. About 30B documents carry pre-computed quality signals and duplicate ids are published, so labs can apply their own filtering thresholds; the deduplicated head and middle buckets total roughly 30T tokens."}
+{"id":"the-pile","name":"The Pile","publisher":"EleutherAI","stage":"pretraining","contentType":"mixed (web + books + papers + code)","tokens":"825GB / 300B","items":"22 sub-corpora","license":"mixed (per-subcorpus)","languages":"English-focused","released":"2020","url":"https://pile.eleuther.ai","notes":"The original open pretraining corpus, now mostly of historical interest. The Books3 subset was taken offline in 2023 after DMCA notices from the Danish Rights Alliance, and the original download host no longer serves the full corpus. EleutherAI released the Common Pile v0.1 in 2025 as its openly licensed successor."}
+{"id":"common-pile-v0.1","name":"Common Pile v0.1","publisher":"EleutherAI (with University of Toronto, Vector Institute, Hugging Face, AI2, and others)","stage":"pretraining","contentType":"mixed (papers, code, books, government text, educational materials, transcripts)","tokens":"8TB","items":"30 sources","license":"public domain and openly licensed text (per-source licenses)","languages":"English-focused","released":"2025-06","url":"https://huggingface.co/common-pile","notes":"Successor to The Pile built only from public-domain and openly licensed text, including an openly licensed subset of The Stack v2 and about 300,000 public-domain books. Released alongside Comma v0.1-1T and Comma v0.1-2T, 7B models trained on 1T and 2T tokens of it."}
+{"id":"dolma","name":"Dolma","publisher":"Allen AI (AI2)","stage":"pretraining","contentType":"web + code + books + papers","tokens":"3T","items":"mixed","license":"ODC-BY-1.0","languages":"English","released":"2024-04","url":"https://huggingface.co/datasets/allenai/dolma","notes":"AI2's open pretraining corpus behind the original OLMo models, with documented source provenance. Relicensed from the AI2 ImpACT license to ODC-BY in April 2024, when v1.7 became the default version. Succeeded by the Dolma 3 mix used for OLMo 3."}
+{"id":"dolma-3-mix","name":"Dolma 3 Mix (6T)","publisher":"Allen AI (AI2)","stage":"pretraining","contentType":"web + papers + code","tokens":"6T","items":null,"license":"ODC-BY-1.0","languages":"English","released":"2025-11","url":"https://huggingface.co/datasets/allenai/dolma3_mix-6T","notes":"The pretraining mix used for Olmo-3-1125-32B, mostly drawn from Common Crawl. AI2 also publishes a 150B-token sample with the same upsampling strategy for smaller experiments."}
+{"id":"fineweb-2","name":"FineWeb2","publisher":"Hugging Face","stage":"pretraining","contentType":"multilingual web text","tokens":"20TB / 3T+ words","items":"5B documents","license":"ODC-BY-1.0","languages":"1000+ languages","released":"2024-12","url":"https://huggingface.co/datasets/HuggingFaceFW/fineweb-2","notes":"Multilingual successor to FineWeb built from 96 CommonCrawl snapshots (2013 to April 2024). The card reports words rather than tokens because token counts vary widely by tokenizer and script. v2.1.0 (June 2025) aligned filtering with the published paper and grew the dataset."}
+{"id":"finepdfs","name":"FinePDFs","publisher":"Hugging Face","stage":"pretraining","contentType":"text extracted from PDFs","tokens":"3T","items":"475M documents","license":"ODC-BY-1.0","languages":"1733 language-script pairs","released":"2025-09","url":"https://huggingface.co/datasets/HuggingFaceFW/finepdfs","notes":"Described by its publisher as the largest public corpus sourced only from PDFs. Fully reproducible, with extraction code released alongside the data."}
+{"id":"finetranslations","name":"FineTranslations","publisher":"Hugging Face","stage":"pretraining","contentType":"parallel text (web text translated into English)","tokens":"1T+","items":null,"license":"ODC-BY-1.0","languages":"English paired with 500+ languages","released":"2026-01","url":"https://huggingface.co/datasets/HuggingFaceFW/finetranslations","notes":"FineWeb2 documents translated into English with Gemma 3 27B using a synthetic-data pipeline in datatrove. An educational-filtered variant is published separately."}
+{"id":"ultra-fineweb-l3","name":"Ultra-FineWeb-L3","publisher":"OpenBMB","stage":"pretraining","contentType":"synthetic web-derived text (Q&A pairs, multi-style rewrites)","tokens":"600B+ (400B+ English, 200B+ Chinese)","items":null,"license":"Apache-2.0","languages":"English, Chinese","released":"2026-02","url":"https://huggingface.co/datasets/openbmb/Ultra-FineWeb-L3","notes":"Refined tier of the UltraData framework, generated from Ultra-FineWeb with MiniCPM4 and Qwen3 and used in the decay phase of MiniCPM5-1B. The L1 filtered tier (1T+ tokens) followed in August 2026."}
+{"id":"refinedweb","name":"RefinedWeb","publisher":"TII","stage":"pretraining","contentType":"web text","tokens":"500-650B (public extract)","items":"968M documents (public extract)","license":"ODC-BY-1.0","languages":"English","released":"2023-06","url":"https://huggingface.co/datasets/tiiuae/falcon-refinedweb","notes":"Behind the Falcon series. Only a public extract of the full corpus was released; use is also subject to the CommonCrawl terms. Aggressive deduplication and filtering; older than FineWeb."}
+{"id":"the-stack-v3","name":"The Stack v3","publisher":"Hugging Face","stage":"pretraining","contentType":"source code (grouped by repository)","tokens":"3.6T (train subset)","items":"173M repositories (train subset)","license":"ODC-BY-1.0 (files keep their original repository licenses)","languages":"713 programming languages (train subset)","released":"2026-07","url":"https://huggingface.co/datasets/HuggingFaceCode/stack-v3-train","notes":"Successor to The Stack v2, crawled from GitHub default branches as of August 2025 with file contents inline. The train subset is 15.9TB; v3.1 removed leaked exact duplicates, cutting it from 4.9T to 3.6T tokens. The full corpus is 113.7TB across 770 languages and 224M repositories, and developers can opt out."}
+{"id":"the-stack-v2","name":"The Stack v2","publisher":"BigCode (HuggingFace + ServiceNow)","stage":"pretraining","contentType":"source code","tokens":"67.5TB full / ~900B (train-full)","items":"3.28B unique files from 104.2M repositories","license":"original repository licenses (gated terms of use)","languages":"600+ programming languages","released":"2024-02","url":"https://huggingface.co/datasets/bigcode/the-stack-v2","notes":"Sourced from the Software Heritage archive and deduplicated at file and near-duplicate level (the dedup version is 32.1TB). Behind StarCoder 2. Access is gated and use must follow the original licenses. Succeeded by The Stack v3 in 2026."}
+{"id":"starcoderdata","name":"StarCoderData","publisher":"BigCode","stage":"pretraining","contentType":"source code + jupyter + GitHub issues + commits","tokens":"~250B","items":"86 languages","license":"mixed permissive","languages":"86 programming languages","released":"2023-05","url":"https://huggingface.co/datasets/bigcode/starcoderdata","notes":"Pre-cleaned subset of The Stack used to train StarCoder: 783GB of code plus GitHub issues, Jupyter notebooks, and commits. Smaller than The Stack v2 and easier to work with for small-scale code-model experiments."}
+{"id":"tulu-3-sft-mix","name":"Tulu 3 SFT Mixture","publisher":"Allen AI","stage":"instruction-tuning","contentType":"instruction-response pairs","tokens":null,"items":"939K instructions","license":"ODC-BY-1.0","languages":"English","released":"2024-11","url":"https://huggingface.co/datasets/allenai/tulu-3-sft-mixture","notes":"SFT data for AI2's Tulu 3 post-training recipe. Combines FLAN v2, OpenAssistant, WildChat, Aya, NuminaMath, persona-generated math, code, and instruction-following prompts, and safety sets. Some subsets carry non-commercial licenses."}
+{"id":"open-hermes-2.5","name":"OpenHermes 2.5","publisher":"Teknium / Nous Research","stage":"instruction-tuning","contentType":"multi-turn conversations","tokens":null,"items":"1M conversations","license":"mixed (per-source)","languages":"English","released":"2023-12","url":"https://huggingface.co/datasets/teknium/OpenHermes-2.5","notes":"Compiled from GPT-4 outputs, Airoboros, ShareGPT, and other open and synthetic sets. The dataset behind OpenHermes 2.5, Nous Hermes 2, and many community fine-tunes."}
+{"id":"openthoughts3","name":"OpenThoughts3-1.2M","publisher":"Open Thoughts","stage":"instruction-tuning","contentType":"reasoning traces (math, code, science)","tokens":null,"items":"1.2M examples (850K math, 250K code, 100K science)","license":"Apache-2.0","languages":"English","released":"2025-06","url":"https://huggingface.co/datasets/open-thoughts/OpenThoughts3-1.2M","notes":"Third OpenThoughts release, built through an ablation-driven pipeline over question sourcing, selection, and answer generation. Used to train OpenThinker3-7B and as part of the SmolLM3 mid-training mix."}
+{"id":"glaive-function-calling","name":"Glaive Function Calling v2","publisher":"Glaive AI","stage":"instruction-tuning","contentType":"function-calling traces","tokens":null,"items":"113K examples","license":"Apache-2.0","languages":"English","released":"2023-12","url":"https://huggingface.co/datasets/glaiveai/glaive-function-calling-v2","notes":"Synthetic function-calling traces. A common starting point for fine-tuning open models on JSON tool use."}
+{"id":"open-orca","name":"OpenOrca","publisher":"Open Orca team","stage":"instruction-tuning","contentType":"reasoning traces","tokens":null,"items":"~1M GPT-4 + ~3.2M GPT-3.5 completions","license":"MIT","languages":"English","released":"2023-07","url":"https://huggingface.co/datasets/Open-Orca/OpenOrca","notes":"Reproduction of the Microsoft Orca paper. Augments FLAN tasks with GPT-4 and GPT-3.5 step-by-step reasoning traces. Strong base for math and reasoning fine-tunes."}
+{"id":"agent-instruct","name":"AgentInstruct (orca-agentinstruct-1M-v1)","publisher":"Microsoft Research","stage":"instruction-tuning","contentType":"synthetic instruction pairs","tokens":null,"items":"~1M instruction pairs (public subset of ~25M)","license":"CDLA-Permissive-2.0","languages":"English","released":"2024-10","url":"https://huggingface.co/datasets/microsoft/orca-agentinstruct-1M-v1","notes":"Fully synthetic data generated with the AgentInstruct agentic framework from public web text seeds: text editing, creative writing, code, reading comprehension, RAG, brain teasers, and more. The full ~25M-pair set was used to post-train Orca-3-Mistral; Microsoft shares the 1M subset for research."}
+{"id":"ultrafeedback","name":"UltraFeedback","publisher":"OpenBMB","stage":"dpo","contentType":"preference pairs","tokens":null,"items":"64K instructions x 4 model responses","license":"MIT","languages":"English","released":"2023-10","url":"https://huggingface.co/datasets/openbmb/UltraFeedback","notes":"Widely used DPO dataset for open models. GPT-4 rates 4 candidate responses per prompt across helpfulness, honesty, instruction-following, and truthfulness."}
+{"id":"tulu-3-pref","name":"Tulu 3 Preference Mixture (70B)","publisher":"Allen AI","stage":"dpo","contentType":"preference pairs","tokens":null,"items":"337K preference pairs","license":"ODC-BY-1.0","languages":"English","released":"2024-11","url":"https://huggingface.co/datasets/allenai/llama-3.1-tulu-3-70b-preference-mixture","notes":"Preference pairs for the Tulu 3 70B DPO stage, with responses generated from a pool of open models including Tulu 2 and InternLM2.5. Some portions carry non-commercial terms."}
+{"id":"helpsteer-2","name":"HelpSteer 2","publisher":"NVIDIA","stage":"dpo","contentType":"preference pairs (5-attribute)","tokens":null,"items":"21K conversations","license":"CC-BY-4.0","languages":"English","released":"2024-06","url":"https://huggingface.co/datasets/nvidia/HelpSteer2","notes":"NVIDIA preference data labeled across helpfulness, correctness, coherence, complexity, verbosity. Strong for multi-attribute reward modeling. Followed by HelpSteer3 in 2025."}
+{"id":"helpsteer-3","name":"HelpSteer3","publisher":"NVIDIA","stage":"dpo","contentType":"preference pairs + feedback + edits","tokens":null,"items":"40K preference samples (plus Feedback and Edit subsets)","license":"CC-BY-4.0","languages":"multilingual (general, STEM, code, and multilingual domains)","released":"2025-03","url":"https://huggingface.co/datasets/nvidia/HelpSteer3","notes":"Human-annotated preferences over responses from about 20 commercially permissive open models, with no outputs from proprietary providers. The Feedback and Edit subsets support inference-time scaling research."}
+{"id":"laion-5b","name":"LAION-5B (Re-LAION-5B)","publisher":"LAION","stage":"multimodal","contentType":"image-text pairs (links + alt text)","tokens":null,"items":"~5.8B image-text pairs","license":"Apache-2.0 (Re-LAION-5B)","languages":"multilingual","released":"2024-08","url":"https://laion.ai/blog/relaion-5b/","notes":"Behind Stable Diffusion and many open image models. The original 2022 release was withdrawn in December 2023 after a Stanford Internet Observatory report found links to CSAM. Re-LAION-5B (August 2024) removed 2,236 flagged links and ships gated on Hugging Face in research and research-safe versions."}
+{"id":"datacomp-1b","name":"DataComp-1B","publisher":"DataComp","stage":"multimodal","contentType":"image-text pairs","tokens":null,"items":"1.4B filtered pairs","license":"CC-BY-4.0 (url-text metadata; images keep their own copyrights)","languages":"multilingual","released":"2023","url":"https://www.datacomp.ai","notes":"Filtered subset of CommonPool. Reproducible filtering recipes are part of the contribution. Behind several open CLIP-class models."}
+{"id":"finevision","name":"FineVision","publisher":"Hugging Face","stage":"multimodal","contentType":"image-text instruction data for vision-language models","tokens":"9.5B answer tokens","items":"24.3M samples, 17.3M images","license":"per sub-dataset (prompts CC-BY-4.0)","languages":"per sub-dataset","released":"2025-09","url":"https://huggingface.co/datasets/HuggingFaceM4/FineVision","notes":"Unifies more than 200 public sources into 185 subsets with 88.9M turns, deduplicated and decontaminated against 66 benchmarks, including agentic and GUI tasks. Each source keeps its own license."}

2026-10-12/training-runs.jsonl ADDED Viewed

	@@ -0,0 +1,14 @@

+{"id":"soofi-s-30b-a3b","model":"Soofi S 30B-A3B","publisher":"Soofi consortium","released":"2026-07","activeParamsB":3,"totalParamsB":30,"trainingTokens":"26.68T","hardware":"NVIDIA B200 (Deutsche Telekom Industrial AI Cloud, Munich)","hardwareCount":512,"computeHours":"~253,000 B200 GPU-hours","estimatedCostMillionUSD":null,"costSource":"estimated","duration":"~7 weeks (March 24 to May 13, 2026)","openWeights":false,"url":"https://arxiv.org/abs/2607.09424","notes":"German-English hybrid Mamba-Transformer MoE built on the Nemotron 3 Nano reference architecture by a German research consortium (Fraunhofer IAIS, DFKI and others). Unusually complete compute disclosure: up to 512 B200s, about 253,000 GPU-hours and the exact training window. Weights are a gated closed-beta preview pending a final license; no dollar cost given."}
+{"id":"longcat-2.0","model":"LongCat-2.0","publisher":"Meituan","released":"2026-07","activeParamsB":48,"totalParamsB":1600,"trainingTokens":"35T+","hardware":"AI ASIC superpods (vendor not named)","hardwareCount":50000,"computeHours":"millions of accelerator-days (per Meituan)","estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":true,"url":"https://huggingface.co/meituan-longcat/LongCat-2.0","notes":"MIT-licensed MoE with about 48B active parameters. Meituan says pretraining ran on more than 50,000 AI ASICs (hardwareCount is that lower bound) with less memory per device than an H800, over 35T+ tokens with no rollbacks, plus hundreds of billions of 1M-context tokens. The accelerator vendor and dollar cost are not disclosed."}
+{"id":"nemotron-3-ultra","model":"Nemotron 3 Ultra","publisher":"NVIDIA","released":"2026-06","activeParamsB":55,"totalParamsB":550,"trainingTokens":"20T","hardware":"NVIDIA GB200 (long-context phase); NVFP4 pretraining","hardwareCount":null,"computeHours":null,"estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":true,"url":"https://research.nvidia.com/labs/nemotron/Nemotron-3-Ultra/","notes":"Hybrid Mamba-Attention MoE pretrained in NVFP4 on 20T text tokens (15T diversity phase, then 5T high-quality phase), then extended to 1M context on GB200 GPUs. NVIDIA calls it the largest demonstration of stable NVFP4 training to date. Base, post-trained and quantized checkpoints plus most training data are on Hugging Face; GPU count and GPU-hours are not disclosed."}
+{"id":"gpt-5.5","model":"GPT-5.5","publisher":"OpenAI","released":"2026-04","activeParamsB":null,"totalParamsB":null,"trainingTokens":"undisclosed","hardware":"NVIDIA GB200 and GB300 NVL72","hardwareCount":null,"computeHours":null,"estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":false,"url":"https://openai.com/index/introducing-gpt-5-5/","notes":"Released April 23, 2026. OpenAI says GPT-5.5 was co-designed for, trained with and served on NVIDIA GB200 and GB300 NVL72 systems. The system card gives no parameter count, token count, cluster size or cost, so no figures are listed here."}
+{"id":"deepseek-v4-pro","model":"DeepSeek V4 Pro","publisher":"DeepSeek","released":"2026-04","activeParamsB":49,"totalParamsB":1600,"trainingTokens":"33T","hardware":"undisclosed (report validates kernels on NVIDIA GPUs and Huawei Ascend NPUs)","hardwareCount":null,"computeHours":null,"estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":true,"url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","notes":"MIT-licensed preview with a 1M-token context. Sibling DeepSeek V4 Flash is 284B total / 13B active on 32T tokens. The technical report details the architecture (hybrid CSA/HCA attention, mHC, Muon optimizer) and data, but unlike the V3 report it gives no GPU count, GPU-hours or training cost."}
+{"id":"claude-opus-4-7","model":"Claude Opus 4.7","publisher":"Anthropic","released":"2026-04","activeParamsB":null,"totalParamsB":null,"trainingTokens":"undisclosed","hardware":"AWS and Google Cloud (accelerators not specified)","hardwareCount":null,"computeHours":null,"estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":false,"url":"https://www.anthropic.com/news/claude-opus-4-7","notes":"Released April 16, 2026. Anthropic's transparency hub lists cloud computing resources from Amazon Web Services and Google Cloud Platform (with PyTorch, JAX and Triton) but no chip types, counts, tokens or cost."}
+{"id":"llama-4-maverick","model":"Llama 4 Maverick","publisher":"Meta","released":"2025-04","activeParamsB":17,"totalParamsB":400,"trainingTokens":"~22T","hardware":"NVIDIA H100-80GB","hardwareCount":null,"computeHours":"2.38M GPU-hours","estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":true,"url":"https://github.com/meta-llama/llama-models/blob/main/models/llama4/MODEL_CARD.md","notes":"Released April 5, 2025. Meta's model card discloses 2.38M H100 GPU-hours of pretraining for Maverick (Llama 4 Scout took 5.0M on about 40T tokens). Cluster size and cost are not disclosed."}
+{"id":"gemini-2.5-pro","model":"Gemini 2.5 Pro","publisher":"Google","released":"2025-03","activeParamsB":null,"totalParamsB":null,"trainingTokens":"undisclosed","hardware":"Google TPUv5p (multiple 8,960-chip pods)","hardwareCount":null,"computeHours":null,"estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":false,"url":"https://arxiv.org/abs/2507.06261","notes":"First announced March 25, 2025. The Gemini 2.5 technical report says the family was the first trained on TPUv5p, with synchronous data-parallel training across multiple 8,960-chip pods spread over several datacenters. No parameter, token or cost figures."}
+{"id":"gpt-4","model":"GPT-4","publisher":"OpenAI","released":"2023-03","activeParamsB":280,"totalParamsB":1760,"trainingTokens":"~13T (estimated)","hardware":"NVIDIA A100","hardwareCount":25000,"computeHours":"~270M GPU-hours (estimated)","estimatedCostMillionUSD":100,"costSource":"estimated","duration":"~3 months","openWeights":false,"url":"https://openai.com/index/gpt-4-research/","notes":"Reverse-engineered numbers from public leaks (1.76T MoE, 280B active). The reference point for \"what did frontier training cost in 2023.\" Made obsolete in compute terms by H100/H200/Blackwell era."}
+{"id":"llama-3.1-405b","model":"Llama 3.1 405B","publisher":"Meta","released":"2024-07","activeParamsB":405,"totalParamsB":405,"trainingTokens":"15T+","hardware":"NVIDIA H100-80GB","hardwareCount":16000,"computeHours":"30.84M GPU-hours","estimatedCostMillionUSD":60,"costSource":"estimated","duration":"undisclosed (paper analyzes a 54-day pretraining window)","openWeights":true,"url":"https://arxiv.org/abs/2407.21783","notes":"Meta's model card documents 30.84M H100 GPU-hours for the 405B model, and the Llama 3 paper says it trained on up to 16K H100s. The ~$60M cost is an estimate at roughly $2 per GPU-hour, not a Meta disclosure."}
+{"id":"deepseek-v3","model":"DeepSeek V3","publisher":"DeepSeek","released":"2024-12","activeParamsB":37,"totalParamsB":671,"trainingTokens":"14.8T","hardware":"NVIDIA H800","hardwareCount":2048,"computeHours":"2.788M GPU-hours","estimatedCostMillionUSD":5.576,"costSource":"disclosed","duration":"under two months (pretraining)","openWeights":true,"url":"https://github.com/deepseek-ai/DeepSeek-V3","notes":"The paper reports 2.664M H800 GPU-hours for pretraining plus 119K for context extension and 5K for post-training on a 2,048-GPU cluster. The $5.576M figure assumes $2 per GPU-hour and covers only the final run, excluding prior research and ablations."}
+{"id":"olmo-2-32b","model":"OLMo 2 32B","publisher":"Allen AI","released":"2025-03","activeParamsB":32,"totalParamsB":32,"trainingTokens":"6.6T","hardware":"NVIDIA H100 (Ai2 Jupiter and Augusta clusters)","hardwareCount":null,"computeHours":null,"estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":true,"url":"https://arxiv.org/abs/2501.00656","notes":"Fully open: weights, training data, code, logs and intermediate checkpoints. The report gives 6.6T total tokens for the 32B model (6.06T in pretraining, the rest in mid-training) and describes Ai2's H100 clusters, but not GPU-hours for this run."}
+{"id":"mistral-large-2","model":"Mistral Large 2","publisher":"Mistral","released":"2024-07","activeParamsB":123,"totalParamsB":123,"trainingTokens":"undisclosed","hardware":"undisclosed","hardwareCount":null,"computeHours":null,"estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":true,"url":"https://mistral.ai/news/mistral-large-2407/","notes":"Released July 24, 2024 under the Mistral Research License: weights are downloadable for research and non-commercial use, and commercial self-deployment needs a separate license. Mistral does not disclose tokens, hardware or cost."}
+{"id":"qwen-2.5-72b","model":"Qwen 2.5 72B","publisher":"Alibaba","released":"2024-09","activeParamsB":72,"totalParamsB":72,"trainingTokens":"18T","hardware":"undisclosed","hardwareCount":null,"computeHours":null,"estimatedCostMillionUSD":null,"costSource":"estimated","duration":"undisclosed","openWeights":true,"url":"https://arxiv.org/abs/2412.15115","notes":"Alibaba's open dense flagship. The Qwen2.5 technical report states 18T pretraining tokens (up from 7T for Qwen2) but gives no hardware, GPU-hours or cost."}

2026-10-12/trending-repos.jsonl ADDED Viewed

	@@ -0,0 +1,20 @@

+{"name":"Significant-Gravitas/AutoGPT","description":"AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matter","language":"Python","stars":187500,"forks":45924,"todayStars":0,"url":"https://github.com/Significant-Gravitas/AutoGPT","topics":["agentic-ai","agents","ai","artificial-intelligence","autonomous-agents","claude","gpt","llama-api","llm","openai","python"],"createdAt":"2023-03-16T09:21:07Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"deepseek-ai/deepseek-harness","description":"DeepSeek Harness: Everything is a Plugin.","language":"TypeScript","stars":163740,"forks":17320,"todayStars":0,"url":"https://github.com/deepseek-ai/deepseek-harness","topics":["ai-agents","cordis","dsh","dsh-plugin"],"createdAt":"2026-08-13T11:56:32Z","fetchedAt":"2026-08-19T08:30:22.199Z"}
+{"name":"langgenius/dify","description":"Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so t","language":"TypeScript","stars":158059,"forks":24955,"todayStars":0,"url":"https://github.com/langgenius/dify","topics":["agent","agentic-ai","agentic-framework","agentic-workflow","ai","automation","claude","deepseek","genai","gpt","llm","low-code","mcp","nextjs","no-code","openai","python","skills","workflow"],"createdAt":"2023-04-12T07:40:24Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"bytedance/deer-flow","description":"An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and ","language":"Python","stars":83619,"forks":11629,"todayStars":0,"url":"https://github.com/bytedance/deer-flow","topics":["agent","agentic","agentic-framework","agentic-workflow","ai","ai-agents","deep-research","harness","langchain","langgraph","langmanus","llm","multi-agent","nodejs","podcast","python","superagent","typescript"],"createdAt":"2025-05-07T02:50:19Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"headroomlabs-ai/headroom","description":"Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same a","language":"Python","stars":74866,"forks":5805,"todayStars":0,"url":"https://github.com/headroomlabs-ai/headroom","topics":["agent","ai","anthropic","claude-code","compression","context-engineering","context-window","cursor","fastapi","langchain","llm","mcp","openai","prompt-engineering","proxy","python","rag","token-optimization","tokens","typescript"],"createdAt":"2026-01-07T19:58:51Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"hesreallyhim/awesome-claude-code","description":"A hand-picked collection of the finest of resources for the most awesome of agents, Claude Code, the undisputed champion of coding companions, from th","language":"Python","stars":55342,"forks":4801,"todayStars":0,"url":"https://github.com/hesreallyhim/awesome-claude-code","topics":["agent-skills","agentic-code","agentic-coding","ai-workflow-optimization","ai-workflows","anthropic","anthropic-claude","awesome","awesome-claude-code","awesome-list","awesome-lists","awesome-resources","claude","claude-code","coding-agent","coding-agents","coding-assistant","coding-assistants","llm"],"createdAt":"2025-04-19T20:55:59Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"oobabooga/textgen","description":"Open-source desktop app for local LLMs. Text, vision, tool-calling, OpenAI/Anthropic-compatible API. 100% private.","language":"Python","stars":47559,"forks":5977,"todayStars":0,"url":"https://github.com/oobabooga/textgen","topics":[],"createdAt":"2022-12-21T04:17:37Z","fetchedAt":"2026-08-22T08:30:51.897Z"}
+{"name":"zhayujie/CowAgent","description":"Open-source personal AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-agent, multi-mode","language":"Python","stars":47310,"forks":10378,"todayStars":0,"url":"https://github.com/zhayujie/CowAgent","topics":["ai","ai-agent","ai-agents","chatgpt-on-wechat","claude","claude-code","codex","cowagent","deepseek","harness","llm","mcp","multi-agent","openai","openclaw","personal-agent","skills"],"createdAt":"2022-08-07T08:33:41Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"oobabooga/text-generation-webui","description":"The original local LLM interface. Text, vision, tool-calling, training, and more. 100% offline.","language":"Python","stars":46462,"forks":5913,"todayStars":0,"url":"https://github.com/oobabooga/text-generation-webui","topics":[],"createdAt":"2022-12-21T04:17:37Z","fetchedAt":"2026-04-11T08:30:32.199Z"}
+{"name":"HKUDS/nanobot","description":"Lightweight, open-source AI agent for your tools, chats, and workflows.","language":"Python","stars":46212,"forks":8178,"todayStars":0,"url":"https://github.com/HKUDS/nanobot","topics":["ai","ai-agent","ai-agents","anthropic","chatgpt","claude","claude-code","codex","codex-cli","llm","nanobot","openai","openclaw"],"createdAt":"2026-02-01T07:16:15Z","fetchedAt":"2026-07-25T08:30:57.424Z"}
+{"name":"alibaba/open-code-review","description":"Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-lev","language":"Go","stars":45595,"forks":3293,"todayStars":0,"url":"https://github.com/alibaba/open-code-review","topics":["agent","agent-skills","code-review","code-review-assistant","harness","repository-level-context"],"createdAt":"2026-05-18T02:16:36Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"milla-jovovich/mempalace","description":"The highest-scoring AI memory system ever benchmarked. And it's free.","language":"Python","stars":40650,"forks":5142,"todayStars":0,"url":"https://github.com/milla-jovovich/mempalace","topics":["ai","chromadb","llm","mcp","memory","python"],"createdAt":"2026-04-05T01:12:07Z","fetchedAt":"2026-04-11T08:30:32.148Z"}
+{"name":"chopratejas/headroom","description":"Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 60-95% fewer tokens, same answers. Library, proxy, MCP server.","language":"Python","stars":39742,"forks":2721,"todayStars":0,"url":"https://github.com/chopratejas/headroom","topics":["agent","ai","anthropic","claude-code","compression","context-engineering","context-window","cursor","fastapi","langchain","llm","mcp","openai","prompt-engineering","proxy","python","rag","token-optimization","tokens","typescript"],"createdAt":"2026-01-07T19:58:51Z","fetchedAt":"2026-06-20T08:30:12.877Z"}
+{"name":"JuliusBrussee/caveman","description":"🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman","language":"Python","stars":36850,"forks":1775,"todayStars":0,"url":"https://github.com/JuliusBrussee/caveman","topics":["ai","anthropic","caveman","claude","claude-code","llm","meme","prompt-engineering","skill","tokens"],"createdAt":"2026-04-04T10:03:00Z","fetchedAt":"2026-04-17T22:15:03.352Z"}
+{"name":"DietrichGebert/ponytail","description":"Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.","language":"JavaScript","stars":33532,"forks":1522,"todayStars":0,"url":"https://github.com/DietrichGebert/ponytail","topics":["agent-skills","ai-agents","claude","claude-code","claude-code-plugin","cursor-rules","developer-tools","llm","prompt-engineering","yagni"],"createdAt":"2026-06-12T00:52:37Z","fetchedAt":"2026-06-18T08:30:39.960Z"}
+{"name":"tirth8205/code-review-graph","description":"Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmar","language":"Python","stars":32015,"forks":2933,"todayStars":0,"url":"https://github.com/tirth8205/code-review-graph","topics":["ai-coding","claude","claude-code","code-review","graphrag","incremental","knowledge-graph","llm","mcp","python","static-analysis","tree-sitter"],"createdAt":"2026-02-26T17:25:13Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"ComposioHQ/composio","description":"Composio powers 1000+ toolkits, tool search, context management, authentication, and a sandboxed workbench to help you build AI agents that turn inten","language":"TypeScript","stars":30480,"forks":4846,"todayStars":0,"url":"https://github.com/ComposioHQ/composio","topics":["agentic-ai","agents","ai","ai-agents","aiagents","developer-tools","function-calling","gpt-4","javascript","js","llm","llmops","mcp","python","remote-mcp-server","sse","typescript"],"createdAt":"2024-02-23T13:58:27Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"safishamsi/graphify","description":"AI coding assistant skill (Claude Code, Codex, OpenCode, Cursor, Gemini CLI, GitHub Copilot CLI, OpenClaw, Factory Droid, Trae, Google Antigravity). T","language":"Python","stars":29077,"forks":3193,"todayStars":0,"url":"https://github.com/safishamsi/graphify","topics":["antigravity","claude-code","codex","gemini","graphrag","knowledge-graph","openclaw","skills"],"createdAt":"2026-04-03T15:49:07Z","fetchedAt":"2026-04-17T22:15:03.352Z"}
+{"name":"yamadashy/repomix","description":"📦 Repomix is a powerful tool that packs your entire repository into a single, AI-friendly file. Perfect for when you need to feed your codebase to La","language":"TypeScript","stars":28780,"forks":1575,"todayStars":0,"url":"https://github.com/yamadashy/repomix","topics":["ai","anthropic","artificial-intelligence","chatbot","chatgpt","claude","deepseek","developer-tools","gemini","genai","generative-ai","gpt","javascript","language-model","llama","llm","mcp","nodejs","openai","typescript"],"createdAt":"2024-07-13T07:11:32Z","fetchedAt":"2026-10-10T08:30:56.710Z"}
+{"name":"can1357/oh-my-pi","description":"⌥  AI Coding agent for the terminal — hash-anchored edits, optimized tool harness, LSP, Python, browser, subagents, and more","language":"TypeScript","stars":26380,"forks":2568,"todayStars":0,"url":"https://github.com/can1357/oh-my-pi","topics":["ai-agent","ai-coding-agent","anthropic","bun","claude","cli","coding-assistant","llm","mcp","multi-provider","openai","rust","terminal","tui","typescript"],"createdAt":"2025-12-31T14:01:28Z","fetchedAt":"2026-08-22T08:30:51.897Z"}

2026-10-12/usage-rankings.jsonl ADDED Viewed

	@@ -0,0 +1,10 @@

+{"rank":1,"model":"DeepSeek V4.1 Flash","provider":"DeepSeek","openrouterId":"deepseek/deepseek-v4.1-flash","tokensB7d":19600,"trend":"up","sharePct":21,"notes":"Released 2026-09-10 as DeepSeek's first Causal Encoder-Decoder model (552B total, 8B active on input, 16B on output), MIT weights. Took the top spot in its first full week, up 24%.","url":"https://api-docs.deepseek.com/news/news260910"}
+{"rank":2,"model":"GLM 5.3 Flash","provider":"Z.ai","openrouterId":"z-ai/glm-5.3-flash","tokensB7d":16300,"trend":"up","sharePct":17.4,"notes":"First natively multimodal GLM-5 model (320B total, 18B active, MIT). Up 16% week over week, second by tokens behind DeepSeek V4.1 Flash.","url":"https://z.ai/blog/glm-5.3-flash"}
+{"rank":3,"model":"Space Bunny Alpha","provider":"Stealth","openrouterId":"stealth/space-bunny-alpha","tokensB7d":13900,"trend":"new","sharePct":14.9,"notes":"Anonymous stealth model OpenRouter listed on 2026-09-23, free for the preview, 1M context with text, image and video input. Reached #3 on five days of traffic.","url":"https://openrouter.ai/stealth/space-bunny-alpha"}
+{"rank":4,"model":"Hy4 preview","provider":"Tencent","openrouterId":"tencent/hy4-preview","tokensB7d":9640,"trend":"down","sharePct":10.3,"notes":"Tencent MoE preview released August 2026: 770B total, 49B active, Apache-2.0 weights. Down 23% week over week as newer flash tier models took share.","url":"https://huggingface.co/tencent/Hy4-preview"}
+{"rank":5,"model":"GPT-5.6 Luna","provider":"OpenAI","openrouterId":"openai/gpt-5.6-luna","tokensB7d":8530,"trend":"down","sharePct":9.1,"notes":"Most cost-efficient GPT-5.6 tier and the highest-volume closed model in the window. Down 12% week over week, with GPT-6 Luna arriving 2026-09-22.","url":"https://openai.com/index/gpt-5-6/"}
+{"rank":6,"model":"DeepSeek V4 Flash 0731","provider":"DeepSeek","openrouterId":"deepseek/deepseek-v4-flash-0731","tokensB7d":7820,"trend":"down","sharePct":8.3,"notes":"Re-post-trained July 31 revision of V4 Flash with MIT weights. Down 17% week over week as DeepSeek traffic moves to V4.1 Flash.","url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731"}
+{"rank":7,"model":"Nemotron 3 Ultra (free)","provider":"NVIDIA","openrouterId":"nvidia/nemotron-3-ultra-550b-a55b:free","tokensB7d":5670,"trend":"up","sharePct":6,"notes":"Free OpenRouter variant of NVIDIA's open 550B MoE (55B active, OpenMDW-1.1 license). Up 26% week over week.","url":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16"}
+{"rank":8,"model":"MiMo-V2.6-Flash","provider":"Xiaomi","openrouterId":"xiaomi/mimo-v2.6-flash","tokensB7d":5510,"trend":"new","sharePct":5.9,"notes":"Xiaomi omnimodal MoE released 2026-09-21 with MIT weights: 309B total, 15B active, 1M context, text, image, audio and video input. Entered at #8, taking the MiMo-V2.5 slot.","url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"}
+{"rank":9,"model":"DeepSeek V4 Flash","provider":"DeepSeek","openrouterId":"deepseek/deepseek-v4-flash","tokensB7d":3350,"trend":"down","sharePct":3.5,"notes":"Original April 2026 checkpoint (OpenRouter label: DeepSeek V4 Flash 0423), 284B total with 13B active, MIT weights. Down 11% week over week.","url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash"}
+{"rank":10,"model":"GPT-6 Luna","provider":"OpenAI","openrouterId":"openai/gpt-6-luna","tokensB7d":2860,"trend":"new","sharePct":3,"notes":"OpenAI's cheapest GPT-6 tier, launched 2026-09-22 at $0.10 input / $0.50 output per 1M with a 1.05M context window. Entered the top 10 on five days of traffic.","url":"https://openai.com/index/introducing-gpt-6-sol-and-luna/"}

2026-10-12/vector-dbs.jsonl ADDED Viewed

	@@ -0,0 +1,14 @@

+{"id":"pinecone","name":"Pinecone","type":"managed","hostingOptions":["cloud"],"freeTier":"Starter: 2GB storage, 5 indexes, 2M write units and 1M read units per month, 100 namespaces per index","startingPriceUSDMonth":20,"pricingNote":"Builder is $20/month flat. Standard carries a $50/month usage minimum (3-week trial with $300 credit) and Enterprise a $500/month minimum. Serverless usage on Standard: $0.33/GB-month storage, $4 to $4.50 per 1M write units, $16 to $18 per 1M read units. Dedicated read nodes and BYOC on the upper tiers.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"unlimited","openSource":false,"license":"Proprietary","released":"2021","url":"https://www.pinecone.io/pricing/","notes":"Most-deployed managed vector DB. Serverless removed instance sizing as a concern, and namespaces give cheap multi-tenancy. Founder Edo Liberty handed the CEO role to Ash Ashutosh in September 2025 amid press reports that the company was weighing a sale; worth tracking for anyone signing a long contract."}
+{"id":"turbopuffer","name":"Turbopuffer","type":"managed","hostingOptions":["cloud"],"freeTier":null,"startingPriceUSDMonth":16,"pricingNote":"Usage-based with monthly minimums: Launch $16, Scale $256 (adds SSO, audit logs, HIPAA BAA), Enterprise from $4,096 with single-tenant and BYOC options. No free tier; per-operation rates come from the pricing calculator.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"unlimited","openSource":false,"license":"Proprietary","released":"2024","url":"https://turbopuffer.com/pricing","notes":"Stores vectors on object storage (S3/GCS) and caches hot data closer to compute, which is where the cost advantage at scale comes from. Cold queries are slower; warm queries compete with in-memory databases. Strong fit for many-tenant workloads with uneven traffic."}
+{"id":"qdrant-cloud","name":"Qdrant Cloud","type":"hybrid","hostingOptions":["cloud","self-host"],"freeTier":"Free forever single-node cluster: 0.5 vCPU, 1GB RAM, 4GB disk","startingPriceUSDMonth":null,"pricingNote":"Standard tier billed on actual compute, memory, and storage usage through a calculator rather than a fixed entry price. Hybrid Cloud and Private Cloud are quoted. Cloud Inference includes some free embedding models. Self-hosted Qdrant is free under Apache-2.0.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":false,"maxVectorsPaid":"unlimited","openSource":true,"license":"Apache-2.0","released":"2021","url":"https://qdrant.tech/pricing/","notes":"Strong open-source story; Apache-2.0 with feature-parity between cloud and self-host. Native sparse vector support (not just bm25 hybrid). Rust core. Serverless is listed as coming soon, so capacity is still sized per cluster."}
+{"id":"weaviate-cloud","name":"Weaviate Cloud","type":"hybrid","hostingOptions":["cloud","self-host"],"freeTier":"Free plan: 100k objects, 1GB memory, 10GB disk, 1 collection","startingPriceUSDMonth":45,"pricingNote":"Flex from $45/month pay-as-you-go on shared clusters (from $0.00465 per 1M vector dimensions, $0.12/GiB storage). Premium from $400/month prepaid with shared or dedicated deployment. Self-hosted Weaviate is free.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"unlimited","openSource":true,"license":"BSD-3-Clause (optional features under the Weaviate License)","released":"2019","url":"https://weaviate.io/pricing","notes":"Schema-first, with first-class hybrid search and built-in modules for embeddings, reranking, and generative answers. License check: since late August 2026 the repo carries a wl directory under a separate Weaviate License, switched on with the WEAVIATE_LICENSE environment variable. With the variable unset, only BSD-3-Clause code runs."}
+{"id":"milvus-zilliz","name":"Milvus / Zilliz Cloud","type":"hybrid","hostingOptions":["cloud","self-host","embedded"],"freeTier":"Zilliz Cloud Free: 5GB storage, 2.5M vCUs per month, up to 5 collections","startingPriceUSDMonth":0,"pricingNote":"Standard from $0/month on serverless with usage billing and a 30-day trial. Enterprise dedicated clusters from $197/month. Dedicated capacity-optimized clusters from $16 per million vectors per month, tiered storage from $5. BYOC available. Milvus self-host free under Apache-2.0.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"tens of billions","openSource":true,"license":"Apache-2.0","released":"2019","url":"https://zilliz.com/pricing","notes":"Most battle-tested at very-large scale (10B+ vectors). GPU index option (CAGRA) for ultra-low-latency ANN. Embedded mode (Milvus Lite) for single-node test harnesses. Zilliz now markets the service as a Vector Lakebase, with on-demand compute jobs over external data."}
+{"id":"chromadb","name":"Chroma","type":"hybrid","hostingOptions":["cloud","self-host","embedded"],"freeTier":"Cloud Starter: $0/month plus usage with $5 free credit. Self-host: free under Apache-2.0","startingPriceUSDMonth":0,"pricingNote":"Chroma Cloud usage-based: $2.50 per logical GiB written, $0.33/GiB-month stored, $0.0075 per TiB queried, $0.09 per GiB returned. Team plan $250/month with $100 credit. Self-host or embedded is free.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":false,"serverless":true,"maxVectorsPaid":"unlimited","openSource":true,"license":"Apache-2.0","released":"2023","url":"https://www.trychroma.com","notes":"Developer-first ergonomics and the default in many LangChain and LlamaIndex tutorials. Chroma Cloud now supports sparse vectors combined with dense embeddings through reciprocal rank fusion, plus copy-on-write collection forking. The embedded library remains the easiest single-process option."}
+{"id":"pgvector","name":"pgvector (Postgres)","type":"oss","hostingOptions":["cloud","self-host","embedded"],"freeTier":"Same as your Postgres host (Supabase Free: 500MB database; Neon Free: 0.5GB per project)","startingPriceUSDMonth":0,"pricingNote":"Pricing follows your Postgres host. Supabase Pro from $25/month. Neon Launch and Scale are pay-as-you-go with no monthly minimum. AWS RDS varies.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"limited by Postgres scaling","openSource":true,"license":"PostgreSQL","released":"2021","url":"https://github.com/pgvector/pgvector","notes":"Postgres extension. Best fit when your app already has Postgres and your vector workload is small to medium. Hybrid via tsvector + vector. HNSW + IVFFlat indexes."}
+{"id":"lancedb","name":"LanceDB","type":"hybrid","hostingOptions":["cloud","self-host","embedded"],"freeTier":"Open-source library and embedded mode free under Apache-2.0","startingPriceUSDMonth":null,"pricingNote":"LanceDB Cloud and Enterprise are priced through sales; no public rate card. The embedded library (in-process Rust with Python and JS bindings) is free.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":false,"serverless":true,"maxVectorsPaid":"unlimited","openSource":true,"license":"Apache-2.0","released":"2023","url":"https://lancedb.com","notes":"Columnar Lance file format on object storage. Embedded mode rivals SQLite for ergonomics: zero-config, in-process. Hybrid search pairs BM25 full-text with vectors and RRF reranking, and analytics queries run on the same table as vector search."}
+{"id":"mongodb-atlas","name":"MongoDB Atlas Vector Search","type":"managed","hostingOptions":["cloud"],"freeTier":"Free cluster (512MB storage), limited to 3 search or vector indexes","startingPriceUSDMonth":8,"pricingNote":"Flex clusters bill $0.011/hour and scale with load from $8 to a $30/month cap, with up to 10 indexes. Dedicated M10 from $0.08/hour. Production vector workloads are pointed at M10+ with dedicated Search Nodes, billed separately.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"unlimited","openSource":false,"license":"Proprietary","released":"2023","url":"https://www.mongodb.com/products/platform/atlas-vector-search","notes":"Best fit when MongoDB is already your operational database. Hybrid via MongoDB Search ($search) plus $vectorSearch in one aggregation pipeline. MongoDB acquired Voyage AI in February 2025, so its embedding and reranking models plug into Atlas directly."}
+{"id":"vespa","name":"Vespa Cloud","type":"hybrid","hostingOptions":["cloud","self-host"],"freeTier":"Free trial","startingPriceUSDMonth":0,"pricingNote":"Usage-based on Vespa Cloud (compute + storage); rates are not published on the pricing page. Self-hosted Vespa free under Apache-2.0.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":false,"maxVectorsPaid":"unlimited","openSource":true,"license":"Apache-2.0","released":"2017","url":"https://vespa.ai","notes":"Yahoo-built; powers very-large-scale search and recommendation systems (10B+ docs). Steepest learning curve in the catalog but the strongest hybrid retrieval and ranking story."}
+{"id":"elasticsearch","name":"Elasticsearch","type":"hybrid","hostingOptions":["cloud","self-host"],"freeTier":"Elastic Cloud free trial; Serverless vector profiles include 50GB free","startingPriceUSDMonth":null,"pricingNote":"Elasticsearch Serverless: search from $0.09 and ingest from $0.14 per VCU-hour, storage from $0.047/GB-month. Elastic Cloud Hosted is resource-based. Self-hosting is free under AGPLv3 or the Elastic License.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"unlimited","openSource":true,"license":"AGPL-3.0 / SSPL / Elastic License 2.0","released":"2022 (vector)","url":"https://www.elastic.co/elasticsearch","notes":"Best fit when you already use Elastic for log or text search and want vectors co-located. Native HNSW with BM25 plus vector hybrid scoring and first-class RRF. Jina AI models are now sold as Elastic models and run natively through the Elastic Inference Service."}
+{"id":"opensearch","name":"OpenSearch","type":"oss","hostingOptions":["cloud","self-host"],"freeTier":null,"startingPriceUSDMonth":0,"pricingNote":"AWS OpenSearch Serverless bills OCU-hours plus storage. NextGen collections have no minimum and scale to zero after 10 idle minutes; Classic collections bill at least 2 OCUs (1 in dev-test mode). Vector collections cannot share OCUs with search collections. Self-host free under Apache-2.0.","hybridSearch":true,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"unlimited","openSource":true,"license":"Apache-2.0","released":"2022 (vector)","url":"https://opensearch.org","notes":"Apache-2.0 fork of Elasticsearch 7.10. AWS-native with serverless option and GPU-accelerated vector indexing on the managed service. Best fit for AWS-heavy stacks that want hybrid search without licensing entanglements."}
+{"id":"s3-vectors","name":"Amazon S3 Vectors","type":"managed","hostingOptions":["cloud"],"freeTier":null,"startingPriceUSDMonth":0,"pricingNote":"Pure usage in US East: $0.06/GB-month storage, $0.20/GB for PUTs, $2.50 per 1M queries plus a per-TB data-processed charge that falls as an index grows past 100K vectors.","hybridSearch":false,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"2 billion per index, 10,000 indexes per bucket","openSource":false,"license":"Proprietary","released":"2025","url":"https://aws.amazon.com/s3/features/vectors/","notes":"Vector storage built into S3, generally available since December 2025. Frequent queries return in about 100ms and infrequent ones in under a second, with up to 50 metadata keys per vector for filtering and 100 results per query. The cheap cold tier for large corpora; pair it with OpenSearch or a hot database when latency or hybrid search matters."}
+{"id":"cloudflare-vectorize","name":"Cloudflare Vectorize","type":"managed","hostingOptions":["cloud"],"freeTier":"Workers Free: 30M queried and 5M stored vector dimensions per month","startingPriceUSDMonth":5,"pricingNote":"Included with Workers Paid ($5/month account plan): 50M queried dimensions per month then $0.01 per 1M, and 10M stored dimensions then $0.05 per 100M. No charge for CPU, memory, index hours, or index count.","hybridSearch":false,"metadataFiltering":true,"multiTenancy":true,"serverless":true,"maxVectorsPaid":"20M per index (up to 1536 dimensions)","openSource":false,"license":"Proprietary","released":"2023","url":"https://developers.cloudflare.com/vectorize/platform/pricing/","notes":"Serverless vector index that lives next to Workers, Workers AI embeddings, and R2, so an agent can embed, store, and query without leaving one platform. Billing is purely vector count times dimensions, which makes cost easy to forecast. Limits to plan around: 1536 dimensions per vector, topK of 50 when returning metadata, and 10 metadata indexes per index. No native keyword search."}

2026-10-12/voice-leaderboards.jsonl ADDED Viewed

	@@ -0,0 +1 @@

+ {"ok":true,"source":"tensorfeed.ai","lastUpdated":"2026-04-30","ttsArena":{"source":"TTS Arena (Hugging Face)","sourceUrl":"https://huggingface.co/spaces/TTS-AGI/TTS-Arena","window":"aggregated public votes","rankings":[{"rank":1,"model":"Eleven v3 (alpha)","provider":"ElevenLabs","elo":1287,"trend":"up","openWeights":false,"notes":"Most expressive; supports audio tags ([whisper], [laughs]). Default for premium voice agents."},{"rank":2,"model":"Cartesia Sonic 2","provider":"Cartesia","elo":1264,"trend":"up","openWeights":false,"notes":"Mamba architecture. Sub-90ms TTFB; fastest in production."},{"rank":3,"model":"Eleven Multilingual v2","provider":"ElevenLabs","elo":1248,"trend":"flat","openWeights":false,"notes":"Older ElevenLabs flagship; still strong on cloning quality."},{"rank":4,"model":"Hailuo 02 Voice","provider":"MiniMax","elo":1232,"trend":"up","openWeights":false,"notes":"Chinese-language strongest; rapidly improving on English."},{"rank":5,"model":"OpenAI TTS-1-HD","provider":"OpenAI","elo":1218,"trend":"flat","openWeights":false,"notes":"Cheap and reliable. 6 fixed voices, no cloning."},{"rank":6,"model":"Deepgram Aura-2","provider":"Deepgram","elo":1205,"trend":"flat","openWeights":false,"notes":"Sub-200ms TTFB. Strong real-time integration with Deepgram Nova STT."},{"rank":7,"model":"PlayHT Dialog 1.0","provider":"PlayHT","elo":1192,"trend":"down","openWeights":false,"notes":"Conversational TTS; turn-taking aware."},{"rank":8,"model":"Kokoro TTS","provider":"community","elo":1178,"trend":"up","openWeights":true,"notes":"82M params, Apache-2.0, runs on CPU. Surprisingly competitive at the small-model tier."},{"rank":9,"model":"Google Cloud TTS Studio","provider":"Google","elo":1163,"trend":"flat","openWeights":false,"notes":"Google studio voices. Strong on enterprise SLA."},{"rank":10,"model":"Fish Audio S1","provider":"Fish Audio","elo":1141,"trend":"down","openWeights":true,"notes":"Open-weights TTS with cloning. Apache-2.0."}]},"asrLeaderboard":{"source":"Open ASR Leaderboard + vendor-published WER","sourceUrl":"https://huggingface.co/spaces/hf-audio/open_asr_leaderboard","benchmark":"Open ASR Leaderboard (LibriSpeech + Common Voice + AMI + GigaSpeech)","rankings":[{"rank":1,"model":"AssemblyAI Universal-2","provider":"AssemblyAI","englishWER":5.6,"multilingualWER":7.8,"rtf":0.04,"openWeights":false,"notes":"Best aggregated WER. Strong on long-form (calls, podcasts)."},{"rank":2,"model":"GPT-4o Transcribe","provider":"OpenAI","englishWER":6.7,"multilingualWER":8.1,"rtf":0.06,"openWeights":false,"notes":"Replaced whisper-1 as OpenAI flagship STT. Lower hallucination than Whisper."},{"rank":3,"model":"Deepgram Nova-3","provider":"Deepgram","englishWER":6.84,"multilingualWER":9.2,"rtf":0.02,"openWeights":false,"notes":"Fastest production STT. 36 languages; strong code-switching."},{"rank":4,"model":"Whisper Large v3","provider":"OpenAI","englishWER":7.4,"multilingualWER":10.5,"rtf":0.18,"openWeights":true,"notes":"Apache-2.0 open weights. Cheapest production STT when hosted on Groq (~$0.0007/min)."},{"rank":5,"model":"Whisper Large v3 Turbo","provider":"OpenAI (community)","englishWER":7.8,"multilingualWER":11.1,"rtf":0.05,"openWeights":true,"notes":"Community-distilled Whisper Turbo. 6x faster than base v3 with similar WER on English."},{"rank":6,"model":"NVIDIA Parakeet TDT 1.1B","provider":"NVIDIA","englishWER":7.1,"multilingualWER":null,"rtf":0.06,"openWeights":true,"notes":"Open Apache-2.0. English-only. Strong WER for an open model."},{"rank":7,"model":"Google Chirp 2","provider":"Google","englishWER":8.2,"multilingualWER":9,"rtf":0.1,"openWeights":false,"notes":"Strongest multilingual coverage (125+ languages). Higher price reflects enterprise SLAs."},{"rank":8,"model":"IBM Watson Speech","provider":"IBM","englishWER":9.1,"multilingualWER":11.3,"rtf":0.12,"openWeights":false,"notes":"Mature enterprise STT; lags in benchmark WER but strong domain customization."}]}}

2026-10-12/x402-adopters.jsonl ADDED Viewed

	@@ -0,0 +1,16 @@

+{"id":"x402-spec","name":"x402 specification","org":"x402 Foundation (Linux Foundation)","category":"spec","status":"spec","networks":["chain-agnostic"],"tokens":["token-agnostic"],"x402Methods":["exact","upto","batch-settlement"],"endpointUrl":null,"websiteUrl":"https://x402.org","repoUrl":"https://github.com/x402-foundation/x402","docsUrl":"https://docs.x402.org","lastVerified":"2026-09-14","notes":"The HTTP-status-402 protocol that revives the long-dormant \"Payment Required\" status code as a machine-payable handshake. Created at Coinbase and contributed to the Linux Foundation, which announced the x402 Foundation on April 2, 2026 and its operational launch on July 14, 2026 with 40 members (premier members include AWS, Cloudflare, Coinbase, Google, Mastercard, Stripe, and Visa). Spec v2 uses CAIP-2 network ids; documented schemes are exact, upto, and batch-settlement, with an auth-capture scheme spec also in the repo. The canonical repo moved from coinbase/x402 to x402-foundation/x402."}
+{"id":"a2a-x402","name":"A2A x402 Extension","org":"Google Agentic Commerce","category":"spec","status":"spec","networks":["chain-agnostic"],"tokens":["token-agnostic"],"x402Methods":["exact"],"endpointUrl":null,"websiteUrl":"https://github.com/google-agentic-commerce/a2a-x402","repoUrl":"https://github.com/google-agentic-commerce/a2a-x402","docsUrl":"https://github.com/google-agentic-commerce/a2a-x402/blob/main/spec/v0.1/spec.md","lastVerified":"2026-09-14","notes":"Extension that carries x402 payments inside the Agent-to-Agent (A2A) protocol: payment-required, payment-submitted, and payment-completed messages let a merchant agent charge a client agent on-chain. Spec v0.1 with a Python library and examples; sits alongside Google's AP2 (Agent Payments Protocol) work in the same org."}
+{"id":"tensorfeed","name":"TensorFeed.ai","org":"TensorFeed.ai","category":"publisher","status":"live","networks":["base","solana"],"tokens":["USDC"],"x402Methods":["exact"],"endpointUrl":"https://tensorfeed.ai/api/premium","websiteUrl":"https://tensorfeed.ai/developers/agent-payments","repoUrl":null,"docsUrl":"https://tensorfeed.ai/.well-known/x402.json","lastVerified":"2026-09-14","notes":"Real-time AI-ecosystem data with paid premium endpoints listed in its x402 manifest, payable on Base or Solana. AFTA-certified: code-enforced no-charge guarantees plus Ed25519-signed receipts on every paid call. End-to-end USDC loop verified on Base mainnet 2026-04-27."}
+{"id":"cdp-facilitator","name":"CDP Facilitator","org":"Coinbase Developer Platform","category":"gateway","status":"live","networks":["base","polygon","arbitrum","world","solana"],"tokens":["USDC"],"x402Methods":["exact","upto","batch-settlement"],"endpointUrl":null,"websiteUrl":"https://docs.cdp.coinbase.com/x402/welcome","repoUrl":"https://github.com/coinbase/cdp-sdk","docsUrl":"https://docs.cdp.coinbase.com/x402/seller/facilitator","lastVerified":"2026-09-14","notes":"Coinbase-hosted facilitator with OFAC and KYT screening on every transaction. Accepts ERC-20 tokens on its EVM networks (all three schemes) and SPL tokens on Solana (exact). First 1,000 on-chain transactions per month free, then $0.001 each; verification is free. Coinbase reports more than 100 million transactions and $28 million in payment volume across Base and Solana. Sellers get indexed in the x402 Bazaar discovery API (15,406 resources listed on 2026-09-14) and can appear in the Agentic Market directory of x402 services."}
+{"id":"aws-waf-ai-monetization","name":"AWS WAF AI traffic monetization","org":"Amazon Web Services","category":"gateway","status":"live","networks":["base","solana"],"tokens":["USDC"],"x402Methods":[],"endpointUrl":null,"websiteUrl":"https://aws.amazon.com/blogs/aws/aws-waf-adds-ai-traffic-monetization-capability-to-help-content-owners-charge-ai-bots-for-content-access/","repoUrl":null,"docsUrl":"https://www.coinbase.com/blog/coinbase-and-aws-let-publishers-accept-agents-as-customers-via-x402","lastVerified":"2026-09-14","notes":"Launched June 15, 2026 as an AWS WAF Bot Control capability for sites behind Amazon CloudFront. A Monetize rule returns HTTP 402 with an x402 price manifest (USDC price, accepted networks such as Base and Solana, wallet address); Coinbase's x402 Facilitator verifies and settles. Pricing per content path, bot category, or verification tier. AWS takes no fee on content revenue; Stripe and MPP support listed as coming soon."}
+{"id":"aws-agentcore-payments","name":"Amazon Bedrock AgentCore Payments","org":"Amazon Web Services","category":"gateway","status":"live","networks":["base","solana"],"tokens":["USDC"],"x402Methods":["upto"],"endpointUrl":null,"websiteUrl":"https://aws.amazon.com/about-aws/whats-new/2026/08/bedrock-agentcore-payments-ga/","repoUrl":null,"docsUrl":"https://docs.aws.amazon.com/bedrock-agentcore/latest/devguide/payments.html","lastVerified":"2026-09-14","notes":"Buyer-side managed payments for agents: preview May 7, 2026, generally available August 18, 2026. When an agent hits HTTP 402, AgentCore negotiates x402, signs with a connected Coinbase CDP wallet or Stripe Privy wallet, and enforces spending limits at the infrastructure layer. GA added the x402 upto scheme for pay-per-inference pricing plus MPP support. A curated Coinbase x402 Bazaar MCP server is exposed through AgentCore Gateway for endpoint discovery."}
+{"id":"cloudflare-monetization-gateway","name":"Cloudflare Monetization Gateway","org":"Cloudflare","category":"gateway","status":"announced","networks":[],"tokens":[],"x402Methods":[],"endpointUrl":null,"websiteUrl":"https://blog.cloudflare.com/monetization-gateway/","repoUrl":null,"docsUrl":null,"lastVerified":"2026-09-14","notes":"Waitlist opened July 1, 2026: charge for any web page, dataset, API, or MCP tool behind Cloudflare, with payment rules enforced at the edge and stablecoin settlement over x402, peer-to-peer to the seller wallet. Supported networks, tokens, and pricing not yet published. Buyer side announced August 4, 2026: Cloudflare Wallets (Account Wallets plus capped Virtual Wallets for agents) and cloudflare.pay handles; only handle reservation is open, wallets due in the coming months."}
+{"id":"payai-facilitator","name":"PayAI Facilitator","org":"PayAI Network","category":"gateway","status":"live","networks":["solana","base","polygon","avalanche","arbitrum","sei","xlayer","skale-base"],"tokens":["token-agnostic"],"x402Methods":["exact","batch-settlement"],"endpointUrl":"https://facilitator.payai.network","websiteUrl":"https://payai.network","repoUrl":null,"docsUrl":"https://docs.payai.network/","lastVerified":"2026-09-14","notes":"Multi-network production facilitator listed in the x402 docs, with no API key required for the free tier. Its live /supported endpoint reports exact on EVM chains and Solana, plus batch-settlement on Solana. Also runs a PayAI Bazaar discovery catalog of x402-payable endpoints and MCP tools."}
+{"id":"pay-sh","name":"Pay.sh","org":"Solana Foundation + Google Cloud","category":"gateway","status":"live","networks":["solana"],"tokens":["stablecoins"],"x402Methods":["exact","upto"],"endpointUrl":null,"websiteUrl":"https://pay.sh","repoUrl":"https://github.com/solana-foundation/pay","docsUrl":"https://pay.sh/docs/llms.txt","lastVerified":"2026-09-14","notes":"Launched May 5, 2026: an API gateway on Google Cloud that lets agents pay per request for Gemini, BigQuery, and other Google Cloud APIs plus community providers, settling in stablecoins on Solana. Speaks both x402 (exact, upto) and MPP. Ships a pay CLI (npm @solana/pay) that wraps curl, Claude Code, and Codex to handle 402 challenges."}
+{"id":"stripe-machine-payments","name":"Stripe machine payments (x402)","org":"Stripe","category":"gateway","status":"beta","networks":["base"],"tokens":["USDC"],"x402Methods":["exact"],"endpointUrl":null,"websiteUrl":"https://docs.stripe.com/payments/machine","repoUrl":"https://github.com/stripe-samples/machine-payments","docsUrl":"https://docs.stripe.com/payments/machine/x402","lastVerified":"2026-09-14","notes":"In preview: a PaymentIntent generates a per-payment Base deposit address, the server returns an x402 exact challenge, and Stripe captures once USDC settles on-chain. Requires the 2026-03-04.preview API version and approval for the Stablecoins and Crypto payment method; only US businesses can accept. Stripe also accepts payments over MPP, the Machine Payments Protocol it co-authored with Tempo. Replaces our earlier \"Link for agents\" listing, whose x402 claim we could not re-verify."}
+{"id":"x402-sdk","name":"x402 SDKs (@x402/*, Python, Go)","org":"x402 Foundation","category":"sdk","status":"sdk","networks":["evm","solana","algorand","aptos","stellar","ton","hedera","keeta"],"tokens":["token-agnostic"],"x402Methods":["exact","upto","batch-settlement"],"endpointUrl":null,"websiteUrl":"https://www.npmjs.com/package/@x402/core","repoUrl":"https://github.com/x402-foundation/x402","docsUrl":"https://docs.x402.org","lastVerified":"2026-09-14","notes":"Reference SDKs now published by the foundation: @x402/core plus per-chain mechanism packages and Express, Hono, Fastify, Next.js, fetch, axios, paywall, and MCP adapters on npm; x402 on PyPI; the Go module under x402-foundation/x402/go/v2. Supersedes the older @coinbase/x402 package (last published December 2025)."}
+{"id":"cloudflare-agents-x402","name":"Cloudflare Agents SDK x402","org":"Cloudflare","category":"sdk","status":"sdk","networks":["base","ethereum-mainnet","polygon","optimism","arbitrum","avalanche","solana","aptos","stellar","sui"],"tokens":["USDC"],"x402Methods":["exact"],"endpointUrl":null,"websiteUrl":"https://developers.cloudflare.com/agents/tools/payments/x402/","repoUrl":"https://github.com/cloudflare/agents","docsUrl":"https://developers.cloudflare.com/agents/tools/payments/x402/","lastVerified":"2026-09-14","notes":"x402 on Workers and the Agents SDK: charge for HTTP content or MCP tools, and pay from agents by wrapping MCP clients with withX402Client (agents/x402). Examples use the public x402.org facilitator."}
+{"id":"tensorfeed-python-sdk","name":"tensorfeed (Python)","org":"TensorFeed.ai","category":"sdk","status":"sdk","networks":["base"],"tokens":["USDC"],"x402Methods":["exact"],"endpointUrl":null,"websiteUrl":"https://pypi.org/project/tensorfeed/","repoUrl":null,"docsUrl":"https://tensorfeed.ai/developers/agent-payments","lastVerified":"2026-09-14","notes":"Python SDK with optional [web3] extra for one-call USDC sign-and-send. Targets the TensorFeed API but the payment helpers are reusable for any AFTA-compliant publisher."}
+{"id":"tensorfeed-js-sdk","name":"tensorfeed (JavaScript)","org":"TensorFeed.ai","category":"sdk","status":"sdk","networks":["base"],"tokens":["USDC"],"x402Methods":["exact"],"endpointUrl":null,"websiteUrl":"https://www.npmjs.com/package/tensorfeed","repoUrl":null,"docsUrl":"https://tensorfeed.ai/developers/agent-payments","lastVerified":"2026-09-14","notes":"TypeScript SDK with full coverage of free + premium endpoints and the x402 confirm flow."}
+{"id":"afta-gateway","name":"afta-gateway","org":"TensorFeed.ai","category":"gateway","status":"gateway","networks":["base"],"tokens":["USDC"],"x402Methods":["exact"],"endpointUrl":null,"websiteUrl":"https://github.com/RipperMercs/afta-gateway","repoUrl":"https://github.com/RipperMercs/afta-gateway","docsUrl":"https://tensorfeed.ai/agent-fair-trade","lastVerified":"2026-09-14","notes":"Drop-in Cloudflare Worker template. Wraps any HTTP API in AFTA primitives: USDC verification, bearer-token credits, Ed25519-signed receipts, code-enforced no-charge ledger. MIT, no protocol fee, no signup. Fork, set 3 secrets, deploy."}
+{"id":"tensorfeed-mcp","name":"tensorfeed-mcp (MCP server)","org":"TensorFeed.ai","category":"reference","status":"live","networks":["base"],"tokens":["USDC"],"x402Methods":["exact"],"endpointUrl":null,"websiteUrl":"https://github.com/RipperMercs/tensorfeed-mcp","repoUrl":"https://github.com/RipperMercs/tensorfeed-mcp","docsUrl":"https://registry.modelcontextprotocol.io/v0.1/servers/ai.tensorfeed%2Fmcp-server/versions/latest","lastVerified":"2026-09-14","notes":"MCP server that consumes a TensorFeed bearer token (paid via USDC on Base) and exposes paid premium tools to any MCP client (Claude Desktop, Claude Code, Cursor, Cline, etc). Reference example of how an MCP server can wrap an x402 publisher."}