Hugging Face's logo

Datasets:
danielrosehill
/
OR-Batch-Inference-Models

Modalities:
Tabular
Text
Formats:
csv
Languages:
English
Size:
< 1K
Tags:
openrouter
llm-pricing
batch-inference
tabular
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
danielrosehill commited on
Commit
fd70523
·
verified ·
1 Parent(s): d9ffe73

Add canonical_slug, hugging_face_id, expiration_date and stable_endpoint label; add data dictionary

Browse files
Files changed (3) hide show
  1. README.md +45 -18
  2. data/batch_models_2026-10-31.csv +72 -72
  3. data_dictionary.json +34 -0
README.md CHANGED
@@ -32,26 +32,53 @@ with batch and sync prices, the providers that serve batch, and a price-tier lab
32
 
33
  Prices change often. Re-check on the day you rely on them.
34
 
35
- ## Columns
36
-
37
- | Column | Meaning |
38
- | --- | --- |
39
- | `snapshot_date` | Date the prices were fetched (UTC) |
40
- | `model_listed_utc` | Date OpenRouter first listed the model (`created`). Not the date batch support arrived |
41
- | `model` / `batch_id` / `name` | Sync slug, batch slug to submit, display name |
42
- | `context_length` | Tokens |
43
- | `input_modalities` | `;`-separated |
44
- | `batch_input_usd_per_m`, `batch_output_usd_per_m` | Batch price per million tokens |
45
- | `sync_input_usd_per_m`, `sync_output_usd_per_m` | Sync price for the same model (empty if not listed) |
46
- | `batch_providers` | `;`-separated providers serving the `:batch` route |
47
- | `tier`, `tier_rank` | `Cheap` (1), `Mid` (2), `Expensive` (3), by batch output price |
48
- | `sub_band` | `1a` ≤ 1.00, `1b` 1.00–3.00, `2a` 3–8, `2b` 8–20, `3a` 20–50, `3b` ≥ 50 (output $/M) |
49
- | `input_band` | `Cheap` < 0.60, `Mid` 0.60–3.00, `Expensive` ≥ 3.00 (input $/M) |
50
- | `input_mismatch` | `True` when `input_band` ≠ `tier` (the model moves tier on input-heavy jobs) |
51
- | `out_in_ratio` | Batch output price ÷ batch input price |
52
- | `reference_job_usd` | Cost of 4.5M input + 0.5M output tokens at batch prices (a bulk-labelling job shape) |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
53
 
54
  Tier boundaries sit in the widest gaps of the 2026-10-31 output-price ladder
55
  (2.50 → 3.75, 7.50 → 10, 15 → 25, 37.50 → 60).
56
 
 
 
 
 
 
 
 
 
 
 
 
 
57
  *Compiled by Claude Opus 5.5 for Daniel Rosehill, 2026-10-31.*
 
32
 
33
  Prices change often. Re-check on the day you rely on them.
34
 
35
+ ## Data dictionary
36
+
37
+ Machine-readable copy: [`data_dictionary.json`](data_dictionary.json). **Origin** says where a value comes from:
38
+
39
+ - **api**: retrieved from OpenRouter's public API at fetch time, reformatted only (units, joins).
40
+ - **enriched**: added by us, either computed from API fields by a stated rule or a judgement label. OpenRouter does not publish these.
41
+
42
+ | Column | Origin | Type | Meaning |
43
+ | --- | --- | --- | --- |
44
+ | `snapshot_date` | enriched | date | UTC date the API was fetched. Same for every row of a file. |
45
+ | `model_listed_utc` | api | date | Date OpenRouter first listed the model (`created`, Unix time to UTC date). Not the date batch support arrived. |
46
+ | `model` | api | string | Sync model slug (`id` with `:batch` removed). |
47
+ | `batch_id` | api | string | Batch slug (`id`) to submit batches with. |
48
+ | `name` | api | string | Display name (`name`). |
49
+ | `context_length` | api | integer | Context window in tokens (`context_length`). |
50
+ | `input_modalities` | api | string | `architecture.input_modalities`, `;`-separated. |
51
+ | `batch_input_usd_per_m` | api | number | Batch input price, US$ per million tokens (`pricing.prompt` x 1e6, base tier). |
52
+ | `batch_output_usd_per_m` | api | number | Batch output price, US$ per million tokens (`pricing.completion` x 1e6, base tier). |
53
+ | `sync_input_usd_per_m` | api | number or null | Input price of the sync id for the same model. Empty if the sync id is not listed. |
54
+ | `sync_output_usd_per_m` | api | number or null | Output price of the sync id for the same model. Empty if not listed. |
55
+ | `batch_providers` | api | string | Providers serving the `:batch` id, from `/api/v1/models/{id}/endpoints`, `;`-separated. |
56
+ | `canonical_slug` | api | string | `canonical_slug`: OpenRouter's permanent slug, usually the model name plus a release-date stamp. |
57
+ | `hugging_face_id` | api | string or null | `hugging_face_id`: weights repo for open-weight models. Empty for closed models. |
58
+ | `expiration_date` | api | date or null | `expiration_date`: announced retirement date. Empty if none announced. |
59
+ | `tier` | enriched | string | Cheap / Mid / Expensive by batch output price: < $3, $3-20, >= $20 per million. Rule: notebook/07-price-tiers.md, scripts/price_tiers.py. |
60
+ | `tier_rank` | enriched | integer | 1 = Cheap, 2 = Mid, 3 = Expensive. |
61
+ | `sub_band` | enriched | string | 1a <= 1.00, 1b 1-3, 2a 3-8, 2b 8-20, 3a 20-50, 3b >= 50 (batch output $/M). |
62
+ | `input_band` | enriched | string | Cheap < 0.60, Mid 0.60-3.00, Expensive >= 3.00 (batch input $/M). |
63
+ | `input_mismatch` | enriched | boolean | True when input_band differs from tier: the model changes tier on input-heavy jobs. |
64
+ | `out_in_ratio` | enriched | number | batch_output_usd_per_m / batch_input_usd_per_m, 1 decimal. |
65
+ | `reference_job_usd` | enriched | number | Cost in US$ of 4.5M input + 0.5M output tokens at batch prices (bulk-labelling job shape, 9:1). |
66
+ | `stable_endpoint` | enriched | boolean | True if the id resolves to one fixed model version. False for dynamic aliases (`~` prefix or `latest` in the id, which OpenRouter says 'always redirects to the latest model') and for closed models whose canonical_slug has no date stamp (an upstream alias the provider can repoint). Rule: scripts/build_hf_dataset.py. |
67
+ | `stability_basis` | enriched | string | Which branch of the stable_endpoint rule applied. |
68
 
69
  Tier boundaries sit in the widest gaps of the 2026-10-31 output-price ladder
70
  (2.50 → 3.75, 7.50 → 10, 15 → 25, 37.50 → 60).
71
 
72
+ ## Endpoint stability
73
+
74
+ `stable_endpoint` says whether the id always resolves to the same model version. OpenRouter's
75
+ dynamic aliases (`~google/gemini-flash-latest`, `~anthropic/claude-sonnet-latest` and 18 others
76
+ on 2026-10-31) silently move to each new release. **None of them has a `:batch` id**, so every
77
+ alias is excluded from batch by construction. Three batch models are still labelled unstable:
78
+ `openai/gpt-4o`, `openai/gpt-4o-mini` and `openai/gpt-4-turbo`. Their `canonical_slug` carries no
79
+ date, so they map to OpenAI's undated names, which OpenAI can repoint to a newer snapshot (it
80
+ did so for `gpt-4o` in 2024). That they are aliases is inferred from the undated slug. OpenRouter
81
+ does not state it. The other 68 have a dated `canonical_slug` (66) or are open weights with a fixed
82
+ Hugging Face repo (2, gpt-oss).
83
+
84
  *Compiled by Claude Opus 5.5 for Daniel Rosehill, 2026-10-31.*
data/batch_models_2026-10-31.csv CHANGED
@@ -1,72 +1,72 @@
1
- snapshot_date,model_listed_utc,model,batch_id,name,context_length,input_modalities,batch_input_usd_per_m,batch_output_usd_per_m,sync_input_usd_per_m,sync_output_usd_per_m,batch_providers,tier,tier_rank,sub_band,input_band,input_mismatch,out_in_ratio,reference_job_usd
2
- 2026-10-31,2025-08-05,openai/gpt-oss-20b,openai/gpt-oss-20b:batch,OpenAI: gpt-oss-20b (batch),131072,text,0.024,0.112,0.018,0.09,DeepInfra,Cheap,1,1a,Cheap,False,4.7,0.16
3
- 2026-10-31,2025-08-05,openai/gpt-oss-120b,openai/gpt-oss-120b:batch,OpenAI: gpt-oss-120b (batch),131072,text,0.0296,0.136,0.15,0.6,DeepInfra,Cheap,1,1a,Cheap,False,4.6,0.2
4
- 2026-10-31,2026-09-10,deepseek/deepseek-v4.1-flash,deepseek/deepseek-v4.1-flash:batch,DeepSeek: DeepSeek V4.1 Flash (batch),1048576,text;image,0.05,0.15,0.045,0.5,Sail Research;DeepInfra,Cheap,1,1a,Cheap,False,3.0,0.3
5
- 2026-10-31,2026-08-26,z-ai/glm-5.3-flash,z-ai/glm-5.3-flash:batch,Z.ai: GLM 5.3 Flash (batch),1048576,text;image;video,0.05,0.18,0.0163,0.1479,Sail Research;DeepInfra,Cheap,1,1a,Cheap,False,3.6,0.32
6
- 2026-10-31,2025-08-07,openai/gpt-5-nano,openai/gpt-5-nano:batch,OpenAI: GPT-5 Nano (batch),400000,text;image;file,0.025,0.2,0.05,0.4,OpenAI,Cheap,1,1a,Cheap,False,8.0,0.21
7
- 2026-10-31,2025-04-14,openai/gpt-4.1-nano,openai/gpt-4.1-nano:batch,OpenAI: GPT-4.1 Nano (batch),1047576,image;text;file,0.05,0.2,0.1,0.4,OpenAI,Cheap,1,1a,Cheap,False,4.0,0.33
8
- 2026-10-31,2026-10-07,anthropic/claude-haiku-5.5,anthropic/claude-haiku-5.5:batch,Anthropic: Claude Haiku 5.5 (batch),1000000,text;image;file,0.05,0.25,0.1,0.5,Anthropic,Cheap,1,1a,Cheap,False,5.0,0.35
9
- 2026-10-31,2026-09-22,openai/gpt-6-luna,openai/gpt-6-luna:batch,OpenAI: GPT-6 Luna (batch),1050000,file;image;text,0.05,0.25,0.1,0.5,OpenAI,Cheap,1,1a,Cheap,False,5.0,0.35
10
- 2026-10-31,2026-09-22,openai/gpt-6-luna-pro,openai/gpt-6-luna-pro:batch,OpenAI: GPT-6 Luna Pro (batch),1050000,file;image;text,0.05,0.25,0.1,0.5,OpenAI,Cheap,1,1a,Cheap,False,5.0,0.35
11
- 2026-10-31,2024-07-18,openai/gpt-4o-mini,openai/gpt-4o-mini:batch,OpenAI: GPT-4o-mini (batch),128000,text;image;file,0.075,0.3,0.15,0.6,OpenAI,Cheap,1,1a,Cheap,False,4.0,0.49
12
- 2026-10-31,2026-07-09,openai/gpt-5.6-luna,openai/gpt-5.6-luna:batch,OpenAI: GPT-5.6 Luna (batch),1050000,file;image;text,0.1,0.6,0.2,1.2,OpenAI,Cheap,1,1a,Cheap,False,6.0,0.75
13
- 2026-10-31,2026-07-09,openai/gpt-5.6-luna-pro,openai/gpt-5.6-luna-pro:batch,OpenAI: GPT-5.6 Luna Pro (batch),1050000,file;image;text,0.1,0.6,0.2,1.2,OpenAI,Cheap,1,1a,Cheap,False,6.0,0.75
14
- 2026-10-31,2026-03-17,openai/gpt-5.4-nano,openai/gpt-5.4-nano:batch,OpenAI: GPT-5.4 Nano (batch),400000,file;image;text,0.1,0.625,0.2,1.25,OpenAI,Cheap,1,1a,Cheap,False,6.2,0.76
15
- 2026-10-31,2026-05-07,google/gemini-3.1-flash-lite,google/gemini-3.1-flash-lite:batch,Google: Gemini 3.1 Flash Lite (batch),1048576,text;image;video;file;audio,0.125,0.75,0.25,1.5,Google,Cheap,1,1a,Cheap,False,6.0,0.94
16
- 2026-10-31,2025-04-14,openai/gpt-4.1-mini,openai/gpt-4.1-mini:batch,OpenAI: GPT-4.1 Mini (batch),1047576,image;text;file,0.2,0.8,0.4,1.6,OpenAI,Cheap,1,1a,Cheap,False,4.0,1.3
17
- 2026-10-31,2026-06-03,qwen/qwen3.7-plus,qwen/qwen3.7-plus:batch,Qwen: Qwen3.7 Plus (batch),262144,text;image,0.23,0.92,0.32,1.28,Alibaba,Cheap,1,1a,Cheap,False,4.0,1.5
18
- 2026-10-31,2025-08-07,openai/gpt-5-mini,openai/gpt-5-mini:batch,OpenAI: GPT-5 Mini (batch),400000,text;image;file,0.125,1.0,0.25,2.0,OpenAI,Cheap,1,1a,Cheap,False,8.0,1.06
19
- 2026-10-31,2026-07-21,google/gemini-3.5-flash-lite,google/gemini-3.5-flash-lite:batch,Google: Gemini 3.5 Flash Lite (batch),1048576,text;image;video;file;audio,0.15,1.25,0.3,2.5,Google,Cheap,1,1b,Cheap,False,8.3,1.3
20
- 2026-10-31,2025-12-17,google/gemini-3-flash-preview,google/gemini-3-flash-preview:batch,Google: Gemini 3 Flash Preview (batch),1048576,text;image;file;audio;video,0.25,1.5,0.5,3.0,Google,Cheap,1,1b,Cheap,False,6.0,1.88
21
- 2026-10-31,2026-07-21,google/gemini-3.6-flash,google/gemini-3.6-flash:batch,Google: Gemini 3.6 Flash (batch),1048576,text;image;video;file;audio,0.375,1.875,0.75,3.75,Google,Cheap,1,1b,Cheap,False,5.0,2.62
22
- 2026-10-31,2026-08-13,google/gemini-3.7-flash,google/gemini-3.7-flash:batch,Google: Gemini 3.7 Flash (batch),1048576,text;image;video;file;audio,0.375,1.875,0.75,3.75,Google,Cheap,1,1b,Cheap,False,5.0,2.62
23
- 2026-10-31,2026-09-02,google/gemini-3.8-flash,google/gemini-3.8-flash:batch,Google: Gemini 3.8 Flash (batch),1048576,text;image;video;file;audio,0.375,1.875,0.825,4.125,Google,Cheap,1,1b,Cheap,False,5.0,2.62
24
- 2026-10-31,2026-08-18,z-ai/glm-5.3,z-ai/glm-5.3:batch,Z.ai: GLM 5.3 (batch),1048576,text,0.45,2.0,0.0699,1.4,DeepInfra,Cheap,1,1b,Cheap,False,4.4,3.02
25
- 2026-10-31,2026-04-30,x-ai/grok-4.3,x-ai/grok-4.3:batch,SpaceXAI: Grok 4.3 (batch),1000000,text;image;file,1.0,2.0,1.25,2.5,xAI,Cheap,1,1b,Mid,True,2.0,5.5
26
- 2026-10-31,2025-01-31,openai/o3-mini,openai/o3-mini:batch,OpenAI: o3 Mini (batch),200000,text;file,0.55,2.2,,,OpenAI,Cheap,1,1b,Cheap,False,4.0,3.58
27
- 2026-10-31,2025-04-16,openai/o4-mini,openai/o4-mini:batch,OpenAI: o4 Mini (batch),200000,image;text;file,0.55,2.2,,,OpenAI,Cheap,1,1b,Cheap,False,4.0,3.58
28
- 2026-10-31,2026-03-17,openai/gpt-5.4-mini,openai/gpt-5.4-mini:batch,OpenAI: GPT-5.4 Mini (batch),400000,file;image;text,0.375,2.25,0.75,4.5,OpenAI,Cheap,1,1b,Cheap,False,6.0,2.81
29
- 2026-10-31,2025-10-15,anthropic/claude-haiku-4.5,anthropic/claude-haiku-4.5:batch,Anthropic: Claude Haiku 4.5 (batch),200000,text;image;file,0.5,2.5,1.0,5.0,Anthropic,Cheap,1,1b,Cheap,False,5.0,3.5
30
- 2026-10-31,2026-10-26,openai/gpt-6.1-terra,openai/gpt-6.1-terra:batch,OpenAI: GPT-6.1 Terra (batch),1050000,file;image;text,0.5,2.5,1.0,5.0,OpenAI,Cheap,1,1b,Cheap,False,5.0,3.5
31
- 2026-10-31,2026-10-26,openai/gpt-6.1-terra-pro,openai/gpt-6.1-terra-pro:batch,OpenAI: GPT-6.1 Terra Pro (batch),1050000,file;image;text,0.5,2.5,1.0,5.0,OpenAI,Cheap,1,1b,Cheap,False,5.0,3.5
32
- 2026-10-31,2026-04-30,mistralai/mistral-medium-3-5,mistralai/mistral-medium-3-5:batch,Mistral: Mistral Medium 3.5 (batch),262144,text;image;file,0.75,3.75,1.5,7.5,Mistral,Mid,2,2a,Mid,False,5.0,5.25
33
- 2026-10-31,2025-04-14,openai/gpt-4.1,openai/gpt-4.1:batch,OpenAI: GPT-4.1 (batch),1047576,image;text;file,1.0,4.0,2.0,8.0,OpenAI,Mid,2,2a,Mid,False,4.0,6.5
34
- 2026-10-31,2025-04-16,openai/o3,openai/o3:batch,OpenAI: o3 (batch),200000,image;text;file,1.0,4.0,2.0,8.0,OpenAI,Mid,2,2a,Mid,False,4.0,6.5
35
- 2026-10-31,2026-05-19,google/gemini-3.5-flash,google/gemini-3.5-flash:batch,Google: Gemini 3.5 Flash (batch),1048576,text;image;video;file;audio,0.75,4.5,1.5,9.0,Google,Mid,2,2a,Mid,False,6.0,5.62
36
- 2026-10-31,2025-08-07,openai/gpt-5,openai/gpt-5:batch,OpenAI: GPT-5 (batch),400000,text;image;file,0.625,5.0,1.25,10.0,OpenAI,Mid,2,2a,Mid,False,8.0,5.31
37
- 2026-10-31,2025-11-13,openai/gpt-5.1,openai/gpt-5.1:batch,OpenAI: GPT-5.1 (batch),400000,image;text;file,0.625,5.0,1.25,10.0,OpenAI,Mid,2,2a,Mid,False,8.0,5.31
38
- 2026-10-31,2026-06-30,anthropic/claude-sonnet-5,anthropic/claude-sonnet-5:batch,Anthropic: Claude Sonnet 5 (batch),1000000,text;image;file,1.0,5.0,2.0,10.0,Anthropic,Mid,2,2a,Mid,False,5.0,7.0
39
- 2026-10-31,2026-09-28,anthropic/claude-sonnet-5.5,anthropic/claude-sonnet-5.5:batch,Anthropic: Claude Sonnet 5.5 (batch),1000000,text;image;file,1.0,5.0,2.0,10.0,Anthropic,Mid,2,2a,Mid,False,5.0,7.0
40
- 2026-10-31,2026-10-13,google/gemini-4-argon,google/gemini-4-argon:batch,Google: Gemini 4 Argon (batch),1048576,audio;file;image;text;video,1.0,5.0,2.0,10.0,Google,Mid,2,2a,Mid,False,5.0,7.0
41
- 2026-10-31,2026-09-22,openai/gpt-6-sol,openai/gpt-6-sol:batch,OpenAI: GPT-6 Sol (batch),1050000,file;image;text,1.0,5.0,2.0,10.0,OpenAI,Mid,2,2a,Mid,False,5.0,7.0
42
- 2026-10-31,2026-09-22,openai/gpt-6-sol-pro,openai/gpt-6-sol-pro:batch,OpenAI: GPT-6 Sol Pro (batch),1050000,file;image;text,1.0,5.0,2.0,10.0,OpenAI,Mid,2,2a,Mid,False,5.0,7.0
43
- 2026-10-31,2026-09-29,openai/gpt-6.1-sol,openai/gpt-6.1-sol:batch,OpenAI: GPT-6.1 Sol (batch),1050000,file;image;text,1.0,5.0,2.0,10.0,OpenAI,Mid,2,2a,Mid,False,5.0,7.0
44
- 2026-10-31,2026-09-29,openai/gpt-6.1-sol-pro,openai/gpt-6.1-sol-pro:batch,OpenAI: GPT-6.1 Sol Pro (batch),1050000,file;image;text,1.0,5.0,2.0,10.0,OpenAI,Mid,2,2a,Mid,False,5.0,7.0
45
- 2026-10-31,2024-05-13,openai/gpt-4o,openai/gpt-4o:batch,OpenAI: GPT-4o (batch),128000,text;image;file,1.25,5.0,2.5,10.0,OpenAI,Mid,2,2a,Mid,False,4.0,8.12
46
- 2026-10-31,2026-02-19,google/gemini-3.1-pro-preview,google/gemini-3.1-pro-preview:batch,Google: Gemini 3.1 Pro Preview (batch),1048576,audio;file;image;text;video,1.0,6.0,2.0,12.0,Google,Mid,2,2a,Mid,False,6.0,7.5
47
- 2026-10-31,2026-07-09,openai/gpt-5.6-terra,openai/gpt-5.6-terra:batch,OpenAI: GPT-5.6 Terra (batch),1050000,file;image;text,1.0,6.0,2.0,12.0,OpenAI,Mid,2,2a,Mid,False,6.0,7.5
48
- 2026-10-31,2026-07-09,openai/gpt-5.6-terra-pro,openai/gpt-5.6-terra-pro:batch,OpenAI: GPT-5.6 Terra Pro (batch),1050000,file;image;text,1.0,6.0,2.0,12.0,OpenAI,Mid,2,2a,Mid,False,6.0,7.5
49
- 2026-10-31,2025-12-10,openai/gpt-5.2,openai/gpt-5.2:batch,OpenAI: GPT-5.2 (batch),400000,file;image;text,0.875,7.0,1.75,14.0,OpenAI,Mid,2,2a,Mid,False,8.0,7.44
50
- 2026-10-31,2026-03-05,openai/gpt-5.4,openai/gpt-5.4:batch,OpenAI: GPT-5.4 (batch),1050000,text;image;file,1.25,7.5,2.5,15.0,OpenAI,Mid,2,2a,Mid,False,6.0,9.38
51
- 2026-10-31,2025-09-29,anthropic/claude-sonnet-4.5,anthropic/claude-sonnet-4.5:batch,Anthropic: Claude Sonnet 4.5 (batch),1000000,text;image;file,1.5,7.5,3.0,15.0,Anthropic,Mid,2,2a,Mid,False,5.0,10.5
52
- 2026-10-31,2026-02-17,anthropic/claude-sonnet-4.6,anthropic/claude-sonnet-4.6:batch,Anthropic: Claude Sonnet 4.6 (batch),1000000,text;image;file,1.5,7.5,3.0,15.0,Anthropic,Mid,2,2a,Mid,False,5.0,10.5
53
- 2026-10-31,2026-09-22,anthropic/claude-opus-5.5,anthropic/claude-opus-5.5:batch,Anthropic: Claude Opus 5.5 (batch),1000000,text;image;file,2.0,10.0,4.0,20.0,Anthropic,Mid,2,2b,Mid,False,5.0,14.0
54
- 2026-10-31,2026-07-09,openai/gpt-5.6-sol,openai/gpt-5.6-sol:batch,OpenAI: GPT-5.6 Sol (batch),1050000,file;image;text,2.0,10.0,4.0,20.0,OpenAI,Mid,2,2b,Mid,False,5.0,14.0
55
- 2026-10-31,2026-07-09,openai/gpt-5.6-sol-pro,openai/gpt-5.6-sol-pro:batch,OpenAI: GPT-5.6 Sol Pro (batch),1050000,file;image;text,2.0,10.0,4.0,20.0,OpenAI,Mid,2,2b,Mid,False,5.0,14.0
56
- 2026-10-31,2026-07-16,moonshotai/kimi-k3,moonshotai/kimi-k3:batch,MoonshotAI: Kimi K3 (batch),1048576,text;image;video,2.28,11.4,0.192,15.0,DeepInfra,Mid,2,2b,Mid,False,5.0,15.96
57
- 2026-10-31,2025-11-24,anthropic/claude-opus-4.5,anthropic/claude-opus-4.5:batch,Anthropic: Claude Opus 4.5 (batch),200000,file;image;text,2.5,12.5,5.0,25.0,Anthropic,Mid,2,2b,Mid,False,5.0,17.5
58
- 2026-10-31,2026-02-04,anthropic/claude-opus-4.6,anthropic/claude-opus-4.6:batch,Anthropic: Claude Opus 4.6 (batch),1000000,text;image;file,2.5,12.5,5.0,25.0,Anthropic,Mid,2,2b,Mid,False,5.0,17.5
59
- 2026-10-31,2026-04-16,anthropic/claude-opus-4.7,anthropic/claude-opus-4.7:batch,Anthropic: Claude Opus 4.7 (batch),1000000,text;image;file,2.5,12.5,5.0,25.0,Anthropic,Mid,2,2b,Mid,False,5.0,17.5
60
- 2026-10-31,2026-05-27,anthropic/claude-opus-4.8,anthropic/claude-opus-4.8:batch,Anthropic: Claude Opus 4.8 (batch),1000000,text;image;file,2.5,12.5,5.0,25.0,Anthropic,Mid,2,2b,Mid,False,5.0,17.5
61
- 2026-10-31,2026-07-24,anthropic/claude-opus-5,anthropic/claude-opus-5:batch,Anthropic: Claude Opus 5 (batch),1000000,text;image;file,2.5,12.5,5.0,25.0,Anthropic,Mid,2,2b,Mid,False,5.0,17.5
62
- 2026-10-31,2026-04-24,openai/gpt-5.5,openai/gpt-5.5:batch,OpenAI: GPT-5.5 (batch),1050000,file;image;text,2.5,15.0,5.0,30.0,OpenAI,Mid,2,2b,Mid,False,6.0,18.75
63
- 2026-10-31,2024-04-09,openai/gpt-4-turbo,openai/gpt-4-turbo:batch,OpenAI: GPT-4 Turbo (batch),128000,text;image,5.0,15.0,,,OpenAI,Mid,2,2b,Expensive,True,3.0,30.0
64
- 2026-10-31,2026-06-09,anthropic/claude-fable-5,anthropic/claude-fable-5:batch,Anthropic: Claude Fable 5 (batch),1000000,text;image;file,5.0,25.0,10.0,50.0,Anthropic,Expensive,3,3a,Expensive,False,5.0,35.0
65
- 2026-10-31,2026-09-01,anthropic/claude-fable-5.1,anthropic/claude-fable-5.1:batch,Anthropic: Claude Fable 5.1 (batch),1000000,text;image;file,5.0,25.0,10.0,50.0,Anthropic,Expensive,3,3a,Expensive,False,5.0,35.0
66
- 2026-10-31,2026-09-04,openai/gpt-6-astra,openai/gpt-6-astra:batch,OpenAI: GPT-6 Astra (batch),1050000,file;image;text,5.0,25.0,10.0,50.0,OpenAI,Expensive,3,3a,Expensive,False,5.0,35.0
67
- 2026-10-31,2026-09-04,openai/gpt-6-astra-pro,openai/gpt-6-astra-pro:batch,OpenAI: GPT-6 Astra Pro (batch),1050000,file;image;text,5.0,25.0,10.0,50.0,OpenAI,Expensive,3,3a,Expensive,False,5.0,35.0
68
- 2026-10-31,2025-08-05,anthropic/claude-opus-4.1,anthropic/claude-opus-4.1:batch,Anthropic: Claude Opus 4.1 (batch),200000,image;text;file,7.5,37.5,15.0,75.0,Anthropic,Expensive,3,3a,Expensive,False,5.0,52.5
69
- 2026-10-31,2025-10-06,openai/gpt-5-pro,openai/gpt-5-pro:batch,OpenAI: GPT-5 Pro (batch),400000,image;text;file,7.5,60.0,15.0,120.0,OpenAI,Expensive,3,3b,Expensive,False,8.0,63.75
70
- 2026-10-31,2025-12-10,openai/gpt-5.2-pro,openai/gpt-5.2-pro:batch,OpenAI: GPT-5.2 Pro (batch),400000,image;text;file,10.5,84.0,21.0,168.0,OpenAI,Expensive,3,3b,Expensive,False,8.0,89.25
71
- 2026-10-31,2026-03-05,openai/gpt-5.4-pro,openai/gpt-5.4-pro:batch,OpenAI: GPT-5.4 Pro (batch),1050000,text;image;file,15.0,90.0,30.0,180.0,OpenAI,Expensive,3,3b,Expensive,False,6.0,112.5
72
- 2026-10-31,2026-04-24,openai/gpt-5.5-pro,openai/gpt-5.5-pro:batch,OpenAI: GPT-5.5 Pro (batch),1050000,file;image;text,15.0,90.0,30.0,180.0,OpenAI,Expensive,3,3b,Expensive,False,6.0,112.5
 
1
+ snapshot_date,model_listed_utc,model,batch_id,name,context_length,input_modalities,batch_input_usd_per_m,batch_output_usd_per_m,sync_input_usd_per_m,sync_output_usd_per_m,batch_providers,canonical_slug,hugging_face_id,expiration_date,tier,tier_rank,sub_band,input_band,input_mismatch,out_in_ratio,reference_job_usd,stable_endpoint,stability_basis
2
+ 2026-10-31,2025-08-05,openai/gpt-oss-20b,openai/gpt-oss-20b:batch,OpenAI: gpt-oss-20b (batch),131072,text,0.024,0.112,0.018,0.09,DeepInfra,openai/gpt-oss-20b,openai/gpt-oss-20b,,Cheap,1,1a,Cheap,False,4.7,0.16,True,"open weights, fixed Hugging Face repo"
3
+ 2026-10-31,2025-08-05,openai/gpt-oss-120b,openai/gpt-oss-120b:batch,OpenAI: gpt-oss-120b (batch),131072,text,0.0296,0.136,0.15,0.6,DeepInfra,openai/gpt-oss-120b,openai/gpt-oss-120b,,Cheap,1,1a,Cheap,False,4.6,0.2,True,"open weights, fixed Hugging Face repo"
4
+ 2026-10-31,2026-09-10,deepseek/deepseek-v4.1-flash,deepseek/deepseek-v4.1-flash:batch,DeepSeek: DeepSeek V4.1 Flash (batch),1048576,text;image,0.05,0.15,0.045,0.5,Sail Research;DeepInfra,deepseek/deepseek-v4.1-flash-20260910,deepseek-ai/DeepSeek-V4.1-Flash,,Cheap,1,1a,Cheap,False,3.0,0.3,True,dated canonical_slug
5
+ 2026-10-31,2026-08-26,z-ai/glm-5.3-flash,z-ai/glm-5.3-flash:batch,Z.ai: GLM 5.3 Flash (batch),1048576,text;image;video,0.05,0.18,0.0163,0.1479,Sail Research;DeepInfra,z-ai/glm-5.3-flash-20260826,zai-org/GLM-5.3-Flash,,Cheap,1,1a,Cheap,False,3.6,0.32,True,dated canonical_slug
6
+ 2026-10-31,2025-08-07,openai/gpt-5-nano,openai/gpt-5-nano:batch,OpenAI: GPT-5 Nano (batch),400000,text;image;file,0.025,0.2,0.05,0.4,OpenAI,openai/gpt-5-nano-2025-08-07,,,Cheap,1,1a,Cheap,False,8.0,0.21,True,dated canonical_slug
7
+ 2026-10-31,2025-04-14,openai/gpt-4.1-nano,openai/gpt-4.1-nano:batch,OpenAI: GPT-4.1 Nano (batch),1047576,image;text;file,0.05,0.2,0.1,0.4,OpenAI,openai/gpt-4.1-nano-2025-04-14,,,Cheap,1,1a,Cheap,False,4.0,0.33,True,dated canonical_slug
8
+ 2026-10-31,2026-10-07,anthropic/claude-haiku-5.5,anthropic/claude-haiku-5.5:batch,Anthropic: Claude Haiku 5.5 (batch),1000000,text;image;file,0.05,0.25,0.1,0.5,Anthropic,anthropic/claude-haiku-5.5-20261007,,,Cheap,1,1a,Cheap,False,5.0,0.35,True,dated canonical_slug
9
+ 2026-10-31,2026-09-22,openai/gpt-6-luna,openai/gpt-6-luna:batch,OpenAI: GPT-6 Luna (batch),1050000,file;image;text,0.05,0.25,0.1,0.5,OpenAI,openai/gpt-6-luna-20260922,,,Cheap,1,1a,Cheap,False,5.0,0.35,True,dated canonical_slug
10
+ 2026-10-31,2026-09-22,openai/gpt-6-luna-pro,openai/gpt-6-luna-pro:batch,OpenAI: GPT-6 Luna Pro (batch),1050000,file;image;text,0.05,0.25,0.1,0.5,OpenAI,openai/gpt-6-luna-pro-20260922,,,Cheap,1,1a,Cheap,False,5.0,0.35,True,dated canonical_slug
11
+ 2026-10-31,2024-07-18,openai/gpt-4o-mini,openai/gpt-4o-mini:batch,OpenAI: GPT-4o-mini (batch),128000,text;image;file,0.075,0.3,0.15,0.6,OpenAI,openai/gpt-4o-mini,,,Cheap,1,1a,Cheap,False,4.0,0.49,False,"undated canonical_slug, closed weights: provider alias that can be repointed"
12
+ 2026-10-31,2026-07-09,openai/gpt-5.6-luna,openai/gpt-5.6-luna:batch,OpenAI: GPT-5.6 Luna (batch),1050000,file;image;text,0.1,0.6,0.2,1.2,OpenAI,openai/gpt-5.6-luna-20260709,,,Cheap,1,1a,Cheap,False,6.0,0.75,True,dated canonical_slug
13
+ 2026-10-31,2026-07-09,openai/gpt-5.6-luna-pro,openai/gpt-5.6-luna-pro:batch,OpenAI: GPT-5.6 Luna Pro (batch),1050000,file;image;text,0.1,0.6,0.2,1.2,OpenAI,openai/gpt-5.6-luna-pro-20260709,,,Cheap,1,1a,Cheap,False,6.0,0.75,True,dated canonical_slug
14
+ 2026-10-31,2026-03-17,openai/gpt-5.4-nano,openai/gpt-5.4-nano:batch,OpenAI: GPT-5.4 Nano (batch),400000,file;image;text,0.1,0.625,0.2,1.25,OpenAI,openai/gpt-5.4-nano-20260317,,,Cheap,1,1a,Cheap,False,6.2,0.76,True,dated canonical_slug
15
+ 2026-10-31,2026-05-07,google/gemini-3.1-flash-lite,google/gemini-3.1-flash-lite:batch,Google: Gemini 3.1 Flash Lite (batch),1048576,text;image;video;file;audio,0.125,0.75,0.25,1.5,Google,google/gemini-3.1-flash-lite-20260507,,,Cheap,1,1a,Cheap,False,6.0,0.94,True,dated canonical_slug
16
+ 2026-10-31,2025-04-14,openai/gpt-4.1-mini,openai/gpt-4.1-mini:batch,OpenAI: GPT-4.1 Mini (batch),1047576,image;text;file,0.2,0.8,0.4,1.6,OpenAI,openai/gpt-4.1-mini-2025-04-14,,,Cheap,1,1a,Cheap,False,4.0,1.3,True,dated canonical_slug
17
+ 2026-10-31,2026-06-03,qwen/qwen3.7-plus,qwen/qwen3.7-plus:batch,Qwen: Qwen3.7 Plus (batch),262144,text;image,0.23,0.92,0.32,1.28,Alibaba,qwen/qwen3.7-plus-20260602,,,Cheap,1,1a,Cheap,False,4.0,1.5,True,dated canonical_slug
18
+ 2026-10-31,2025-08-07,openai/gpt-5-mini,openai/gpt-5-mini:batch,OpenAI: GPT-5 Mini (batch),400000,text;image;file,0.125,1.0,0.25,2.0,OpenAI,openai/gpt-5-mini-2025-08-07,,,Cheap,1,1a,Cheap,False,8.0,1.06,True,dated canonical_slug
19
+ 2026-10-31,2026-07-21,google/gemini-3.5-flash-lite,google/gemini-3.5-flash-lite:batch,Google: Gemini 3.5 Flash Lite (batch),1048576,text;image;video;file;audio,0.15,1.25,0.3,2.5,Google,google/gemini-3.5-flash-lite-20260721,,,Cheap,1,1b,Cheap,False,8.3,1.3,True,dated canonical_slug
20
+ 2026-10-31,2025-12-17,google/gemini-3-flash-preview,google/gemini-3-flash-preview:batch,Google: Gemini 3 Flash Preview (batch),1048576,text;image;file;audio;video,0.25,1.5,0.5,3.0,Google,google/gemini-3-flash-preview-20251217,,,Cheap,1,1b,Cheap,False,6.0,1.88,True,dated canonical_slug
21
+ 2026-10-31,2026-07-21,google/gemini-3.6-flash,google/gemini-3.6-flash:batch,Google: Gemini 3.6 Flash (batch),1048576,text;image;video;file;audio,0.375,1.875,0.75,3.75,Google,google/gemini-3.6-flash-20260721,,,Cheap,1,1b,Cheap,False,5.0,2.62,True,dated canonical_slug
22
+ 2026-10-31,2026-08-13,google/gemini-3.7-flash,google/gemini-3.7-flash:batch,Google: Gemini 3.7 Flash (batch),1048576,text;image;video;file;audio,0.375,1.875,0.75,3.75,Google,google/gemini-3.7-flash-20260813,,,Cheap,1,1b,Cheap,False,5.0,2.62,True,dated canonical_slug
23
+ 2026-10-31,2026-09-02,google/gemini-3.8-flash,google/gemini-3.8-flash:batch,Google: Gemini 3.8 Flash (batch),1048576,text;image;video;file;audio,0.375,1.875,0.825,4.125,Google,google/gemini-3.8-flash-20260902,,,Cheap,1,1b,Cheap,False,5.0,2.62,True,dated canonical_slug
24
+ 2026-10-31,2026-08-18,z-ai/glm-5.3,z-ai/glm-5.3:batch,Z.ai: GLM 5.3 (batch),1048576,text,0.45,2.0,0.0699,1.4,DeepInfra,z-ai/glm-5.3-20260816,zai-org/GLM-5.3,,Cheap,1,1b,Cheap,False,4.4,3.02,True,dated canonical_slug
25
+ 2026-10-31,2026-04-30,x-ai/grok-4.3,x-ai/grok-4.3:batch,SpaceXAI: Grok 4.3 (batch),1000000,text;image;file,1.0,2.0,1.25,2.5,xAI,x-ai/grok-4.3-20260430,,,Cheap,1,1b,Mid,True,2.0,5.5,True,dated canonical_slug
26
+ 2026-10-31,2025-01-31,openai/o3-mini,openai/o3-mini:batch,OpenAI: o3 Mini (batch),200000,text;file,0.55,2.2,,,OpenAI,openai/o3-mini-2025-01-31,,,Cheap,1,1b,Cheap,False,4.0,3.58,True,dated canonical_slug
27
+ 2026-10-31,2025-04-16,openai/o4-mini,openai/o4-mini:batch,OpenAI: o4 Mini (batch),200000,image;text;file,0.55,2.2,,,OpenAI,openai/o4-mini-2025-04-16,,,Cheap,1,1b,Cheap,False,4.0,3.58,True,dated canonical_slug
28
+ 2026-10-31,2026-03-17,openai/gpt-5.4-mini,openai/gpt-5.4-mini:batch,OpenAI: GPT-5.4 Mini (batch),400000,file;image;text,0.375,2.25,0.75,4.5,OpenAI,openai/gpt-5.4-mini-20260317,,,Cheap,1,1b,Cheap,False,6.0,2.81,True,dated canonical_slug
29
+ 2026-10-31,2025-10-15,anthropic/claude-haiku-4.5,anthropic/claude-haiku-4.5:batch,Anthropic: Claude Haiku 4.5 (batch),200000,text;image;file,0.5,2.5,1.0,5.0,Anthropic,anthropic/claude-4.5-haiku-20251001,,,Cheap,1,1b,Cheap,False,5.0,3.5,True,dated canonical_slug
30
+ 2026-10-31,2026-10-26,openai/gpt-6.1-terra,openai/gpt-6.1-terra:batch,OpenAI: GPT-6.1 Terra (batch),1050000,file;image;text,0.5,2.5,1.0,5.0,OpenAI,openai/gpt-6.1-terra-20261026,,,Cheap,1,1b,Cheap,False,5.0,3.5,True,dated canonical_slug
31
+ 2026-10-31,2026-10-26,openai/gpt-6.1-terra-pro,openai/gpt-6.1-terra-pro:batch,OpenAI: GPT-6.1 Terra Pro (batch),1050000,file;image;text,0.5,2.5,1.0,5.0,OpenAI,openai/gpt-6.1-terra-pro-20261026,,,Cheap,1,1b,Cheap,False,5.0,3.5,True,dated canonical_slug
32
+ 2026-10-31,2026-04-30,mistralai/mistral-medium-3-5,mistralai/mistral-medium-3-5:batch,Mistral: Mistral Medium 3.5 (batch),262144,text;image;file,0.75,3.75,1.5,7.5,Mistral,mistralai/mistral-medium-3.5-20260430,,,Mid,2,2a,Mid,False,5.0,5.25,True,dated canonical_slug
33
+ 2026-10-31,2025-04-14,openai/gpt-4.1,openai/gpt-4.1:batch,OpenAI: GPT-4.1 (batch),1047576,image;text;file,1.0,4.0,2.0,8.0,OpenAI,openai/gpt-4.1-2025-04-14,,,Mid,2,2a,Mid,False,4.0,6.5,True,dated canonical_slug
34
+ 2026-10-31,2025-04-16,openai/o3,openai/o3:batch,OpenAI: o3 (batch),200000,image;text;file,1.0,4.0,2.0,8.0,OpenAI,openai/o3-2025-04-16,,,Mid,2,2a,Mid,False,4.0,6.5,True,dated canonical_slug
35
+ 2026-10-31,2026-05-19,google/gemini-3.5-flash,google/gemini-3.5-flash:batch,Google: Gemini 3.5 Flash (batch),1048576,text;image;video;file;audio,0.75,4.5,1.5,9.0,Google,google/gemini-3.5-flash-20260519,,,Mid,2,2a,Mid,False,6.0,5.62,True,dated canonical_slug
36
+ 2026-10-31,2025-08-07,openai/gpt-5,openai/gpt-5:batch,OpenAI: GPT-5 (batch),400000,text;image;file,0.625,5.0,1.25,10.0,OpenAI,openai/gpt-5-2025-08-07,,,Mid,2,2a,Mid,False,8.0,5.31,True,dated canonical_slug
37
+ 2026-10-31,2025-11-13,openai/gpt-5.1,openai/gpt-5.1:batch,OpenAI: GPT-5.1 (batch),400000,image;text;file,0.625,5.0,1.25,10.0,OpenAI,openai/gpt-5.1-20251113,,,Mid,2,2a,Mid,False,8.0,5.31,True,dated canonical_slug
38
+ 2026-10-31,2026-06-30,anthropic/claude-sonnet-5,anthropic/claude-sonnet-5:batch,Anthropic: Claude Sonnet 5 (batch),1000000,text;image;file,1.0,5.0,2.0,10.0,Anthropic,anthropic/claude-sonnet-5-20260630,,,Mid,2,2a,Mid,False,5.0,7.0,True,dated canonical_slug
39
+ 2026-10-31,2026-09-28,anthropic/claude-sonnet-5.5,anthropic/claude-sonnet-5.5:batch,Anthropic: Claude Sonnet 5.5 (batch),1000000,text;image;file,1.0,5.0,2.0,10.0,Anthropic,anthropic/claude-sonnet-5.5-20260928,,,Mid,2,2a,Mid,False,5.0,7.0,True,dated canonical_slug
40
+ 2026-10-31,2026-10-13,google/gemini-4-argon,google/gemini-4-argon:batch,Google: Gemini 4 Argon (batch),1048576,audio;file;image;text;video,1.0,5.0,2.0,10.0,Google,google/gemini-4-argon-20261013,,,Mid,2,2a,Mid,False,5.0,7.0,True,dated canonical_slug
41
+ 2026-10-31,2026-09-22,openai/gpt-6-sol,openai/gpt-6-sol:batch,OpenAI: GPT-6 Sol (batch),1050000,file;image;text,1.0,5.0,2.0,10.0,OpenAI,openai/gpt-6-sol-20260922,,,Mid,2,2a,Mid,False,5.0,7.0,True,dated canonical_slug
42
+ 2026-10-31,2026-09-22,openai/gpt-6-sol-pro,openai/gpt-6-sol-pro:batch,OpenAI: GPT-6 Sol Pro (batch),1050000,file;image;text,1.0,5.0,2.0,10.0,OpenAI,openai/gpt-6-sol-pro-20260922,,,Mid,2,2a,Mid,False,5.0,7.0,True,dated canonical_slug
43
+ 2026-10-31,2026-09-29,openai/gpt-6.1-sol,openai/gpt-6.1-sol:batch,OpenAI: GPT-6.1 Sol (batch),1050000,file;image;text,1.0,5.0,2.0,10.0,OpenAI,openai/gpt-6.1-sol-20260929,,,Mid,2,2a,Mid,False,5.0,7.0,True,dated canonical_slug
44
+ 2026-10-31,2026-09-29,openai/gpt-6.1-sol-pro,openai/gpt-6.1-sol-pro:batch,OpenAI: GPT-6.1 Sol Pro (batch),1050000,file;image;text,1.0,5.0,2.0,10.0,OpenAI,openai/gpt-6.1-sol-pro-20260929,,,Mid,2,2a,Mid,False,5.0,7.0,True,dated canonical_slug
45
+ 2026-10-31,2024-05-13,openai/gpt-4o,openai/gpt-4o:batch,OpenAI: GPT-4o (batch),128000,text;image;file,1.25,5.0,2.5,10.0,OpenAI,openai/gpt-4o,,,Mid,2,2a,Mid,False,4.0,8.12,False,"undated canonical_slug, closed weights: provider alias that can be repointed"
46
+ 2026-10-31,2026-02-19,google/gemini-3.1-pro-preview,google/gemini-3.1-pro-preview:batch,Google: Gemini 3.1 Pro Preview (batch),1048576,audio;file;image;text;video,1.0,6.0,2.0,12.0,Google,google/gemini-3.1-pro-preview-20260219,,,Mid,2,2a,Mid,False,6.0,7.5,True,dated canonical_slug
47
+ 2026-10-31,2026-07-09,openai/gpt-5.6-terra,openai/gpt-5.6-terra:batch,OpenAI: GPT-5.6 Terra (batch),1050000,file;image;text,1.0,6.0,2.0,12.0,OpenAI,openai/gpt-5.6-terra-20260709,,,Mid,2,2a,Mid,False,6.0,7.5,True,dated canonical_slug
48
+ 2026-10-31,2026-07-09,openai/gpt-5.6-terra-pro,openai/gpt-5.6-terra-pro:batch,OpenAI: GPT-5.6 Terra Pro (batch),1050000,file;image;text,1.0,6.0,2.0,12.0,OpenAI,openai/gpt-5.6-terra-pro-20260709,,,Mid,2,2a,Mid,False,6.0,7.5,True,dated canonical_slug
49
+ 2026-10-31,2025-12-10,openai/gpt-5.2,openai/gpt-5.2:batch,OpenAI: GPT-5.2 (batch),400000,file;image;text,0.875,7.0,1.75,14.0,OpenAI,openai/gpt-5.2-20251211,,,Mid,2,2a,Mid,False,8.0,7.44,True,dated canonical_slug
50
+ 2026-10-31,2026-03-05,openai/gpt-5.4,openai/gpt-5.4:batch,OpenAI: GPT-5.4 (batch),1050000,text;image;file,1.25,7.5,2.5,15.0,OpenAI,openai/gpt-5.4-20260305,,,Mid,2,2a,Mid,False,6.0,9.38,True,dated canonical_slug
51
+ 2026-10-31,2025-09-29,anthropic/claude-sonnet-4.5,anthropic/claude-sonnet-4.5:batch,Anthropic: Claude Sonnet 4.5 (batch),1000000,text;image;file,1.5,7.5,3.0,15.0,Anthropic,anthropic/claude-4.5-sonnet-20250929,,,Mid,2,2a,Mid,False,5.0,10.5,True,dated canonical_slug
52
+ 2026-10-31,2026-02-17,anthropic/claude-sonnet-4.6,anthropic/claude-sonnet-4.6:batch,Anthropic: Claude Sonnet 4.6 (batch),1000000,text;image;file,1.5,7.5,3.0,15.0,Anthropic,anthropic/claude-4.6-sonnet-20260217,,,Mid,2,2a,Mid,False,5.0,10.5,True,dated canonical_slug
53
+ 2026-10-31,2026-09-22,anthropic/claude-opus-5.5,anthropic/claude-opus-5.5:batch,Anthropic: Claude Opus 5.5 (batch),1000000,text;image;file,2.0,10.0,4.0,20.0,Anthropic,anthropic/claude-opus-5.5-20260921,,,Mid,2,2b,Mid,False,5.0,14.0,True,dated canonical_slug
54
+ 2026-10-31,2026-07-09,openai/gpt-5.6-sol,openai/gpt-5.6-sol:batch,OpenAI: GPT-5.6 Sol (batch),1050000,file;image;text,2.0,10.0,4.0,20.0,OpenAI,openai/gpt-5.6-sol-20260709,,,Mid,2,2b,Mid,False,5.0,14.0,True,dated canonical_slug
55
+ 2026-10-31,2026-07-09,openai/gpt-5.6-sol-pro,openai/gpt-5.6-sol-pro:batch,OpenAI: GPT-5.6 Sol Pro (batch),1050000,file;image;text,2.0,10.0,4.0,20.0,OpenAI,openai/gpt-5.6-sol-pro-20260709,,,Mid,2,2b,Mid,False,5.0,14.0,True,dated canonical_slug
56
+ 2026-10-31,2026-07-16,moonshotai/kimi-k3,moonshotai/kimi-k3:batch,MoonshotAI: Kimi K3 (batch),1048576,text;image;video,2.28,11.4,0.192,15.0,DeepInfra,moonshotai/kimi-k3-20260715,moonshotai/Kimi-K3,,Mid,2,2b,Mid,False,5.0,15.96,True,dated canonical_slug
57
+ 2026-10-31,2025-11-24,anthropic/claude-opus-4.5,anthropic/claude-opus-4.5:batch,Anthropic: Claude Opus 4.5 (batch),200000,file;image;text,2.5,12.5,5.0,25.0,Anthropic,anthropic/claude-4.5-opus-20251124,,,Mid,2,2b,Mid,False,5.0,17.5,True,dated canonical_slug
58
+ 2026-10-31,2026-02-04,anthropic/claude-opus-4.6,anthropic/claude-opus-4.6:batch,Anthropic: Claude Opus 4.6 (batch),1000000,text;image;file,2.5,12.5,5.0,25.0,Anthropic,anthropic/claude-4.6-opus-20260205,,,Mid,2,2b,Mid,False,5.0,17.5,True,dated canonical_slug
59
+ 2026-10-31,2026-04-16,anthropic/claude-opus-4.7,anthropic/claude-opus-4.7:batch,Anthropic: Claude Opus 4.7 (batch),1000000,text;image;file,2.5,12.5,5.0,25.0,Anthropic,anthropic/claude-4.7-opus-20260416,,,Mid,2,2b,Mid,False,5.0,17.5,True,dated canonical_slug
60
+ 2026-10-31,2026-05-27,anthropic/claude-opus-4.8,anthropic/claude-opus-4.8:batch,Anthropic: Claude Opus 4.8 (batch),1000000,text;image;file,2.5,12.5,5.0,25.0,Anthropic,anthropic/claude-4.8-opus-20260528,,,Mid,2,2b,Mid,False,5.0,17.5,True,dated canonical_slug
61
+ 2026-10-31,2026-07-24,anthropic/claude-opus-5,anthropic/claude-opus-5:batch,Anthropic: Claude Opus 5 (batch),1000000,text;image;file,2.5,12.5,5.0,25.0,Anthropic,anthropic/claude-opus-5-20260723,,,Mid,2,2b,Mid,False,5.0,17.5,True,dated canonical_slug
62
+ 2026-10-31,2026-04-24,openai/gpt-5.5,openai/gpt-5.5:batch,OpenAI: GPT-5.5 (batch),1050000,file;image;text,2.5,15.0,5.0,30.0,OpenAI,openai/gpt-5.5-20260423,,,Mid,2,2b,Mid,False,6.0,18.75,True,dated canonical_slug
63
+ 2026-10-31,2024-04-09,openai/gpt-4-turbo,openai/gpt-4-turbo:batch,OpenAI: GPT-4 Turbo (batch),128000,text;image,5.0,15.0,,,OpenAI,openai/gpt-4-turbo,,,Mid,2,2b,Expensive,True,3.0,30.0,False,"undated canonical_slug, closed weights: provider alias that can be repointed"
64
+ 2026-10-31,2026-06-09,anthropic/claude-fable-5,anthropic/claude-fable-5:batch,Anthropic: Claude Fable 5 (batch),1000000,text;image;file,5.0,25.0,10.0,50.0,Anthropic,anthropic/claude-5-fable-20260609,,,Expensive,3,3a,Expensive,False,5.0,35.0,True,dated canonical_slug
65
+ 2026-10-31,2026-09-01,anthropic/claude-fable-5.1,anthropic/claude-fable-5.1:batch,Anthropic: Claude Fable 5.1 (batch),1000000,text;image;file,5.0,25.0,10.0,50.0,Anthropic,anthropic/claude-fable-5.1-20260831,,,Expensive,3,3a,Expensive,False,5.0,35.0,True,dated canonical_slug
66
+ 2026-10-31,2026-09-04,openai/gpt-6-astra,openai/gpt-6-astra:batch,OpenAI: GPT-6 Astra (batch),1050000,file;image;text,5.0,25.0,10.0,50.0,OpenAI,openai/gpt-6-astra-20260903,,,Expensive,3,3a,Expensive,False,5.0,35.0,True,dated canonical_slug
67
+ 2026-10-31,2026-09-04,openai/gpt-6-astra-pro,openai/gpt-6-astra-pro:batch,OpenAI: GPT-6 Astra Pro (batch),1050000,file;image;text,5.0,25.0,10.0,50.0,OpenAI,openai/gpt-6-astra-pro-20260903,,,Expensive,3,3a,Expensive,False,5.0,35.0,True,dated canonical_slug
68
+ 2026-10-31,2025-08-05,anthropic/claude-opus-4.1,anthropic/claude-opus-4.1:batch,Anthropic: Claude Opus 4.1 (batch),200000,image;text;file,7.5,37.5,15.0,75.0,Anthropic,anthropic/claude-4.1-opus-20250805,,,Expensive,3,3a,Expensive,False,5.0,52.5,True,dated canonical_slug
69
+ 2026-10-31,2025-10-06,openai/gpt-5-pro,openai/gpt-5-pro:batch,OpenAI: GPT-5 Pro (batch),400000,image;text;file,7.5,60.0,15.0,120.0,OpenAI,openai/gpt-5-pro-2025-10-06,,,Expensive,3,3b,Expensive,False,8.0,63.75,True,dated canonical_slug
70
+ 2026-10-31,2025-12-10,openai/gpt-5.2-pro,openai/gpt-5.2-pro:batch,OpenAI: GPT-5.2 Pro (batch),400000,image;text;file,10.5,84.0,21.0,168.0,OpenAI,openai/gpt-5.2-pro-20251211,,,Expensive,3,3b,Expensive,False,8.0,89.25,True,dated canonical_slug
71
+ 2026-10-31,2026-03-05,openai/gpt-5.4-pro,openai/gpt-5.4-pro:batch,OpenAI: GPT-5.4 Pro (batch),1050000,text;image;file,15.0,90.0,30.0,180.0,OpenAI,openai/gpt-5.4-pro-20260305,,,Expensive,3,3b,Expensive,False,6.0,112.5,True,dated canonical_slug
72
+ 2026-10-31,2026-04-24,openai/gpt-5.5-pro,openai/gpt-5.5-pro:batch,OpenAI: GPT-5.5 Pro (batch),1050000,file;image;text,15.0,90.0,30.0,180.0,OpenAI,openai/gpt-5.5-pro-20260423,,,Expensive,3,3b,Expensive,False,6.0,112.5,True,dated canonical_slug
data_dictionary.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": "danielrosehill/OR-Batch-Inference-Models",
3
+ "updated": "2026-10-31",
4
+ "origin_values": {
5
+ "api": "Retrieved from OpenRouter's public API (https://openrouter.ai/api/v1/models or /api/v1/models/{id}/endpoints) at fetch time, reformatted only (units, joins).",
6
+ "enriched": "Added by us: computed from API fields by a stated rule, or a judgement label. Not something OpenRouter publishes."
7
+ },
8
+ "columns": [
9
+ {"name": "snapshot_date", "type": "date", "origin": "enriched", "description": "UTC date the API was fetched. Same for every row of a file."},
10
+ {"name": "model_listed_utc", "type": "date", "origin": "api", "description": "Date OpenRouter first listed the model (`created`, Unix time to UTC date). Not the date batch support arrived."},
11
+ {"name": "model", "type": "string", "origin": "api", "description": "Sync model slug (`id` with `:batch` removed)."},
12
+ {"name": "batch_id", "type": "string", "origin": "api", "description": "Batch slug (`id`) to submit batches with."},
13
+ {"name": "name", "type": "string", "origin": "api", "description": "Display name (`name`)."},
14
+ {"name": "context_length", "type": "integer", "origin": "api", "description": "Context window in tokens (`context_length`)."},
15
+ {"name": "input_modalities", "type": "string", "origin": "api", "description": "`architecture.input_modalities`, `;`-separated."},
16
+ {"name": "batch_input_usd_per_m", "type": "number", "origin": "api", "description": "Batch input price, US$ per million tokens (`pricing.prompt` x 1e6, base tier)."},
17
+ {"name": "batch_output_usd_per_m", "type": "number", "origin": "api", "description": "Batch output price, US$ per million tokens (`pricing.completion` x 1e6, base tier)."},
18
+ {"name": "sync_input_usd_per_m", "type": "number|null", "origin": "api", "description": "Input price of the sync id for the same model. Empty if the sync id is not listed."},
19
+ {"name": "sync_output_usd_per_m", "type": "number|null", "origin": "api", "description": "Output price of the sync id for the same model. Empty if not listed."},
20
+ {"name": "batch_providers", "type": "string", "origin": "api", "description": "Providers serving the `:batch` id, from `/api/v1/models/{id}/endpoints`, `;`-separated."},
21
+ {"name": "canonical_slug", "type": "string", "origin": "api", "description": "`canonical_slug`: OpenRouter's permanent slug, usually the model name plus a release-date stamp."},
22
+ {"name": "hugging_face_id", "type": "string|null", "origin": "api", "description": "`hugging_face_id`: weights repo for open-weight models. Empty for closed models."},
23
+ {"name": "expiration_date", "type": "date|null", "origin": "api", "description": "`expiration_date`: announced retirement date. Empty if none announced."},
24
+ {"name": "tier", "type": "string", "origin": "enriched", "description": "Cheap / Mid / Expensive by batch output price: < $3, $3-20, >= $20 per million. Rule: notebook/07-price-tiers.md, scripts/price_tiers.py."},
25
+ {"name": "tier_rank", "type": "integer", "origin": "enriched", "description": "1 = Cheap, 2 = Mid, 3 = Expensive."},
26
+ {"name": "sub_band", "type": "string", "origin": "enriched", "description": "1a <= 1.00, 1b 1-3, 2a 3-8, 2b 8-20, 3a 20-50, 3b >= 50 (batch output $/M)."},
27
+ {"name": "input_band", "type": "string", "origin": "enriched", "description": "Cheap < 0.60, Mid 0.60-3.00, Expensive >= 3.00 (batch input $/M)."},
28
+ {"name": "input_mismatch", "type": "boolean", "origin": "enriched", "description": "True when input_band differs from tier: the model changes tier on input-heavy jobs."},
29
+ {"name": "out_in_ratio", "type": "number", "origin": "enriched", "description": "batch_output_usd_per_m / batch_input_usd_per_m, 1 decimal."},
30
+ {"name": "reference_job_usd", "type": "number", "origin": "enriched", "description": "Cost in US$ of 4.5M input + 0.5M output tokens at batch prices (bulk-labelling job shape, 9:1)."},
31
+ {"name": "stable_endpoint", "type": "boolean", "origin": "enriched", "description": "True if the id resolves to one fixed model version. False for dynamic aliases (`~` prefix or `latest` in the id, which OpenRouter says 'always redirects to the latest model') and for closed models whose canonical_slug has no date stamp (an upstream alias the provider can repoint). Rule: scripts/build_hf_dataset.py."},
32
+ {"name": "stability_basis", "type": "string", "origin": "enriched", "description": "Which branch of the stable_endpoint rule applied."}
33
+ ]
34
+ }