| { |
| "4": { |
| "inputs": { |
| "image": "android.jpg" |
| }, |
| "class_type": "LoadImage", |
| "_meta": { |
| "title": "Load Image" |
| } |
| }, |
| "42": { |
| "inputs": { |
| "text": "Anthropic launches Claude Fable 5, its first public Mythos-class model\nAnthropic on Tuesday launched Claude Fable 5, making a Mythos-class AI model publicly available for the first time — a move the company had promised since restricting its most powerful system to a handful of cyber defenders in April.\n\nThe model shares its underlying architecture with Claude Mythos 5, which remains limited to organizations approved through Anthropic's Project Glasswing program. The difference lies in the guardrails: when Fable 5's classifiers detect queries related to cybersecurity, biology and chemistry, or model distillation, the system automatically routes the response to Claude Opus 4.8 instead. Anthropic says more than 95% of sessions involve no such fallback.\n\nPublished\n16 hours ago\n\nsources\nInside Apple's Siri revamp, Anthropic readies Mythos\nsecurityweek.com\nWhat Fable 5 Can Do\nAnthropic describes Fable 5 as exceeding any model it has previously made generally available, with particular strengths in software engineering, knowledge work, and vision. Early testing partners have reported striking results: Stripe said the model compressed months of engineering work into days, performing a codebase-wide migration in a single day across 50 million lines of Ruby code that would have taken a team over two months.\n\nThe model also beat Pokémon FireRed using only a minimal vision-based harness — a task that had stumped previous Claude models even with additional scaffolding tools. On financial reasoning benchmarks, Fable 5 achieved the highest score of any model tested.\n\n\n1 source\nSafety Trade-offs and Data Retention\nThe release comes with a mandatory 30-day data retention policy on all traffic for Mythos-class models, applicable to both first- and third-party surfaces. Anthropic says the retained data will not be used for training and will be deleted after the 30-day window in most cases.\n\nAnthropic acknowledged the safeguards are \"stricter than would be ideal,\" noting that benign requests sometimes trigger the classifiers. An external bug bounty produced no universal jailbreaks in over 1,000 hours of testing, though the UK AI Safety Institute made progress toward one during initial evaluations.\n\n\n1 source\nPricing and Access\nBoth Fable 5 and Mythos 5 are priced at $10 per million input tokens and $50 per million output tokens. For subscription users on Pro, Max, Team, and Enterprise plans, Fable 5 is available at no extra cost through June 22, after which usage credits will be required.\n\nClaude Mythos 5 — the same model with cyber safeguards lifted — is available immediately to existing Glasswing partners, with Anthropic planning a broader trusted access program in consultation with the US government. A separate biology-focused trusted access track will enroll researchers needing the model's chemistry and biology capabilities without restrictions.\n\nTech journalist Alex Heath first reported the imminent release on Sunday, writing that Anthropic was \"planning to release a public version of Mythos\" with \"substantial guardrails.\"" |
| }, |
| "class_type": "TextInput_", |
| "_meta": { |
| "title": "Text Input ♾️Mixlab" |
| } |
| }, |
| "78": { |
| "inputs": { |
| "filename_prefix": [ |
| "66:95", |
| 0 |
| ], |
| "format": "mp4", |
| "codec": "h264", |
| "video": [ |
| "66:91", |
| 0 |
| ] |
| }, |
| "class_type": "SaveVideo", |
| "_meta": { |
| "title": "Save Video" |
| } |
| }, |
| "66:2": { |
| "inputs": { |
| "seed": 139311448793553, |
| "width": 512, |
| "height": 512, |
| "pipeline": [ |
| "66:6", |
| 0 |
| ], |
| "ref_audio": [ |
| "66:70", |
| 0 |
| ], |
| "avatar_image": [ |
| "66:9", |
| 0 |
| ] |
| }, |
| "class_type": "RunningHub SoulX-FlashHead Sampler", |
| "_meta": { |
| "title": "RunningHub SoulX-FlashHead Sampler" |
| } |
| }, |
| "66:43": { |
| "inputs": { |
| "delimiter": ", ", |
| "clean_whitespace": "true", |
| "text_a": [ |
| "66:40", |
| 0 |
| ], |
| "text_b": [ |
| "42", |
| 0 |
| ], |
| "text_c": [ |
| "66:41", |
| 0 |
| ] |
| }, |
| "class_type": "Text Concatenate", |
| "_meta": { |
| "title": "Text Concatenate" |
| } |
| }, |
| "66:9": { |
| "inputs": { |
| "width": 512, |
| "height": 512, |
| "upscale_method": "nearest-exact", |
| "keep_proportion": "crop", |
| "pad_color": "0, 0, 0", |
| "crop_position": "center", |
| "divisible_by": 2, |
| "device": "gpu", |
| "image": [ |
| "4", |
| 0 |
| ] |
| }, |
| "class_type": "ImageResizeKJv2", |
| "_meta": { |
| "title": "Resize Image v2" |
| } |
| }, |
| "66:32": { |
| "inputs": { |
| "system": [ |
| "66:43", |
| 0 |
| ], |
| "prompt": "do not use **. ", |
| "think": true, |
| "format": "text", |
| "reset_session": true, |
| "connectivity": [ |
| "66:34", |
| 0 |
| ] |
| }, |
| "class_type": "OllamaChat", |
| "_meta": { |
| "title": "Ollama Chat (POST /api/chat)" |
| } |
| }, |
| "66:34": { |
| "inputs": { |
| "url": "http://127.0.0.1:11434", |
| "model": "litert-community/gemma-4-E2B-it-litert-lm:latest", |
| "keep_alive": 5, |
| "keep_alive_unit": "minutes", |
| "🔄 Reconnect": null |
| }, |
| "class_type": "OllamaConnectivityV2", |
| "_meta": { |
| "title": "Ollama Connectivity" |
| } |
| }, |
| "66:67": { |
| "inputs": { |
| "device": "cuda", |
| "audio": [ |
| "66:77", |
| 0 |
| ] |
| }, |
| "class_type": "ResembleDenoise", |
| "_meta": { |
| "title": "Resemble Denoise" |
| } |
| }, |
| "66:70": { |
| "inputs": { |
| "solver": "euler", |
| "nfe": 64, |
| "lambd": 0.1, |
| "tau": 0.5, |
| "device": "cuda", |
| "audio": [ |
| "66:67", |
| 0 |
| ] |
| }, |
| "class_type": "ResembleEnhance", |
| "_meta": { |
| "title": "Resemble Enhance" |
| } |
| }, |
| "66:40": { |
| "inputs": { |
| "text": "You are a professional news anchor.\nTranslate the following news article into English and format it as a natural, engaging script for a television news broadcast.\n\n[Rules]\n- Write the script in clear, spoken English, suitable for reading aloud.\n- Use a professional yet conversational tone, typical of international news broadcasts.\n- Begin the script with a natural introductory phrase (e.g., \"Good evening. Today, we have an important update regarding...\").\n- Rephrase complex technical terms into plain English, providing brief supplementary explanations if necessary.\n- Ensure all proper nouns are clearly transliterated into English.\n- Convert numbers and currencies to standard English formats.\n- Keep sentences concise and conversational to maintain a natural speaking rhythm.\n- DO NOT output any section headers (such as \"Introduction:\", \"Background:\", etc.). Write a single continuous, flowing monologue.\n- Use smooth, spoken transitions between the background context, specific details, and future outlook.\n- Output ONLY the spoken words of the script; do not include any extra explanations or annotations.\n\n[Input Article]" |
| }, |
| "class_type": "TextInput_", |
| "_meta": { |
| "title": "Text Input ♾️Mixlab" |
| } |
| }, |
| "66:41": { |
| "inputs": { |
| "text": "[Output Format]\nWrite the script as a continuous spoken monologue without any section labels. The speech should flow naturally in this order:\n1. Introduction (greeting and breaking the news)\n2. Background (context of the event)\n3. Details (specifics of what happened)\n4. Future (what to expect next)\n5. Sign-off (conclude the broadcast by explicitly saying a closing phrase such as \"This has been brought to you by The Android Times.\" or \"Reporting for The Android Times.\")" |
| }, |
| "class_type": "TextInput_", |
| "_meta": { |
| "title": "Text Input ♾️Mixlab" |
| } |
| }, |
| "66:6": { |
| "inputs": { |
| "model_type": "lite" |
| }, |
| "class_type": "RunningHub SoulX-FlashHead Loader", |
| "_meta": { |
| "title": "RunningHub SoulX-FlashHead Loader" |
| } |
| }, |
| "66:73": { |
| "inputs": { |
| "audio": "android.wav", |
| "audioUI": "/api/view?filename=android.wav&type=input&subfolder=&rand=0.48531220558090404" |
| }, |
| "class_type": "LoadAudio", |
| "_meta": { |
| "title": "Load Audio" |
| } |
| }, |
| "66:33": { |
| "inputs": { |
| "text_0": "Good evening. Today, we have some major news in the world of artificial intelligence as Anthropic has just made a significant model publicly available for the first time. They have launched Claude Fable 5, which is their first Mythos-class AI model released to the public.\n\nThis move fulfills a promise the company had made earlier this year when they restricted their most powerful systems to a select group of cyber defenders. While Fable 5 shares the underlying architecture with Anthropic’s highly secure Claude Mythos 5, it operates with different safety measures built in. Specifically, the system includes automatic classifiers that route sensitive queries—those related to cybersecurity, biology, chemistry, or model distillation—to a more restricted version of the model, Claude Opus 4.8. Anthropic reports that over ninety-five percent of sessions involve no such fallback, indicating that the safety mechanisms are largely working as intended.\n\nWhat is exciting about Fable 5 are its capabilities. Anthropic describes this new model as surpassing any other publicly available model in certain areas, showing particular strengths in software engineering, knowledge work, and vision. Early testing has yielded remarkable results; for example, one partner reported that the model was able to compress months of complex engineering work into just a single day, handling a massive codebase migration across fifty million lines of Ruby code in one session. Furthermore, Fable 5 demonstrated impressive visual reasoning, even beating previous Claude models on tasks involving vision-based harnesses, and achieved the highest score on financial reasoning benchmarks among all tested models.\n\nOf course, this release comes with important considerations regarding safety and access. Anthropic has implemented a mandatory thirty-day data retention policy for all traffic related to these Mythos-class models. They state that this retained data will not be used for future training and will be deleted after the window closes. While Anthropic acknowledges that their safeguards are stricter than they might have ideally been, sometimes benign requests trigger these classifiers. Despite ongoing external testing, no universal jailbreaks have been found over a thousand hours of evaluation.\n\nRegarding access, both Fable 5 and Mythos 5 are priced at ten dollars per million input tokens and fifty dollars per million output tokens. For subscription users on Pro, Max, Team, and Enterprise plans, Fable 5 is currently available at no extra cost through June twenty-second, after which usage credits will be required. Additionally, the full Claude Mythos 5 model, with those advanced cyber safeguards lifted, is immediately accessible to existing Glasswing partners, while Anthropic plans a broader trusted access program in coordination with the United States government.\n\nThis development signals a clear push toward democratizing access to highly capable AI systems while maintaining stringent safety protocols behind the scenes.\n\nReporting for The Android Times.", |
| "text": [ |
| "66:32", |
| 0 |
| ] |
| }, |
| "class_type": "ShowText|pysssss", |
| "_meta": { |
| "title": "Show Text 🐍" |
| } |
| }, |
| "66:77": { |
| "inputs": { |
| "version": "v1.5", |
| "top_k": 30, |
| "top_p": 0.8, |
| "temperature": 1, |
| "num_beams": 3, |
| "max_mel_tokens": 1000, |
| "max_text_tokens_per_sentence": 120, |
| "sentences_bucket_max_size": 4, |
| "fast_inference": true, |
| "custom_cuda_kernel": false, |
| "unload_model": true, |
| "emo_alpha": 1, |
| "audio": [ |
| "66:73", |
| 0 |
| ], |
| "text": [ |
| "66:33", |
| 0 |
| ] |
| }, |
| "class_type": "IndexTTSRun", |
| "_meta": { |
| "title": "IndexTTS Run" |
| } |
| }, |
| "66:88": { |
| "inputs": { |
| "Video": [ |
| "66:2", |
| 0 |
| ] |
| }, |
| "class_type": "Video2ImgConvertor", |
| "_meta": { |
| "title": "Video2ImgConvertor" |
| } |
| }, |
| "66:89": { |
| "inputs": { |
| "facedetection": "retinaface_resnet50", |
| "model": "codeformer-v0.1.0.pth", |
| "visibility": 0.5, |
| "codeformer_weight": 0.5, |
| "image": [ |
| "66:88", |
| 0 |
| ] |
| }, |
| "class_type": "ReActorRestoreFace", |
| "_meta": { |
| "title": "Restore Face 🌌 ReActor" |
| } |
| }, |
| "66:90": { |
| "inputs": { |
| "ckpt_name": "rife47.pth", |
| "clear_cache_after_n_frames": 10, |
| "multiplier": 2, |
| "fast_mode": true, |
| "ensemble": true, |
| "scale_factor": 1, |
| "frames": [ |
| "66:89", |
| 0 |
| ] |
| }, |
| "class_type": "RIFE VFI", |
| "_meta": { |
| "title": "RIFE VFI (recommend rife47 and rife49)" |
| } |
| }, |
| "66:91": { |
| "inputs": { |
| "IMAGE": [ |
| "66:90", |
| 0 |
| ] |
| }, |
| "class_type": "Img2VideoConvertor", |
| "_meta": { |
| "title": "Img2VideoConvertor" |
| } |
| }, |
| "66:92": { |
| "inputs": { |
| "url": "http://127.0.0.1:11434", |
| "model": "litert-community/gemma-4-E2B-it-litert-lm:latest", |
| "keep_alive": 5, |
| "keep_alive_unit": "minutes", |
| "🔄 Reconnect": null |
| }, |
| "class_type": "OllamaConnectivityV2", |
| "_meta": { |
| "title": "Ollama Connectivity" |
| } |
| }, |
| "66:93": { |
| "inputs": { |
| "text": [ |
| "42", |
| 0 |
| ] |
| }, |
| "class_type": "TextInput_", |
| "_meta": { |
| "title": "Text Input ♾️Mixlab" |
| } |
| }, |
| "66:94": { |
| "inputs": { |
| "delimiter": "\\n\\n", |
| "clean_whitespace": "true", |
| "text_a": [ |
| "66:93", |
| 0 |
| ] |
| }, |
| "class_type": "Text Concatenate", |
| "_meta": { |
| "title": "Text Concatenate" |
| } |
| }, |
| "66:95": { |
| "inputs": { |
| "system": [ |
| "66:94", |
| 0 |
| ], |
| "prompt": "do not use **. ", |
| "think": true, |
| "format": "text", |
| "reset_session": true, |
| "connectivity": [ |
| "66:92", |
| 0 |
| ] |
| }, |
| "class_type": "OllamaChat", |
| "_meta": { |
| "title": "Ollama Chat" |
| } |
| } |
| } |
