Skip to content

Listings

All listings: OpenRouter

Every listing for the models on page 1 of 26 of OpenRouter, including each model's other versions and routes. Back to OpenRouter

Sold by OpenRouter

298 listings on this page sold by OpenRouter.

Select a score for details.

About these columns
llminfo IQ
llminfo IQ (capability index). How llminfo IQ is computed
Coding IQ
llminfo Coding IQ (coding capability, its own scale). How Coding IQ is computed
Value
Value for money. 50 means as expected for its results. Above 50, cheaper per task. How Value for money works
Yap
Tokens used per coding task vs. models with the same results. Includes agents that write code. Chattiness on maths doesn't predict it. How the Yap score works
Modelllminfo IQCoding IQInput $/1MOutput $/1MMax Input TokensValueYapEpoch Capabilities Index
Gemini 4 Neon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 18 Nov 2026
150.3
From 88 benchmarks. Plenty of data behind it.
≈ 126 ±31
Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$3.00cache read $0.15$15.001M
49.5
Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
60 ±9
Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Gemini 4 Neon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 18 Nov 2026
google/gemini-4-neon:batch · batch
150.3
From 88 benchmarks. Plenty of data behind it.
≈ 126 ±31
Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.50cache read $0.075$7.501M
49.5
Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
60 ±9
Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Gemini 4 Neon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 18 Nov 2026
Inference by Google
150.3
From 88 benchmarks. Plenty of data behind it.
≈ 126 ±31
Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$3.00cache read $0.15$15.001M
49.5
Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
60 ±9
Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Gemini 4 Neon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 18 Nov 2026
google/gemini-4-neon:batch · batch+global
Inference by Google
150.3
From 88 benchmarks. Plenty of data behind it.
≈ 126 ±31
Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.50cache read $0.075$7.501M
49.5
Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
60 ±9
Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Gemini 4 Neon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 18 Nov 2026
flex
Inference by Google
150.3
From 88 benchmarks. Plenty of data behind it.
≈ 126 ±31
Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.50cache read $0.075$7.501M
49.5
Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
60 ±9
Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Gemini 4 Neon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 18 Nov 2026
global
Inference by Google
150.3
From 88 benchmarks. Plenty of data behind it.
≈ 126 ±31
Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$3.00cache read $0.15$15.001M
49.5
Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
60 ±9
Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Gemini 4 Neon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 18 Nov 2026
global/priority
Inference by Google
150.3
From 88 benchmarks. Plenty of data behind it.
≈ 126 ±31
Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.40cache read $0.27$27.001M
49.5
Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
60 ±9
Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Gemini 4 Neon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 18 Nov 2026
priority
Inference by Google
150.3
From 88 benchmarks. Plenty of data behind it.
≈ 126 ±31
Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.40cache read $0.27$27.001M
49.5
Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
60 ±9
Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Gemini 4 Carbon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 10 Nov 2026
149.4
From 127 benchmarks. Plenty of data behind it.
144 ±11
Coding IQ 143.6, 95% interval 133–155. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.001M
72.1
Good value (72.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
169.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Carbon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 10 Nov 2026
google/gemini-4-carbon:batch · batch
149.4
From 127 benchmarks. Plenty of data behind it.
144 ±11
Coding IQ 143.6, 95% interval 133–155. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.001M
72.1
Good value (72.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
169.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Carbon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 10 Nov 2026
Inference by Google
149.4
From 127 benchmarks. Plenty of data behind it.
144 ±11
Coding IQ 143.6, 95% interval 133–155. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.001M
72.1
Good value (72.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
169.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Carbon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 10 Nov 2026
google/gemini-4-carbon:batch · batch+global
Inference by Google
149.4
From 127 benchmarks. Plenty of data behind it.
144 ±11
Coding IQ 143.6, 95% interval 133–155. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.001M
72.1
Good value (72.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
169.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Carbon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 10 Nov 2026
flex
Inference by Google
149.4
From 127 benchmarks. Plenty of data behind it.
144 ±11
Coding IQ 143.6, 95% interval 133–155. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.001M
72.1
Good value (72.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
169.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Carbon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 10 Nov 2026
global/priority
Inference by Google
149.4
From 127 benchmarks. Plenty of data behind it.
144 ±11
Coding IQ 143.6, 95% interval 133–155. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$3.60cache read $0.18$18.001M
72.1
Good value (72.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
169.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Carbon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 10 Nov 2026
priority
Inference by Google
149.4
From 127 benchmarks. Plenty of data behind it.
144 ±11
Coding IQ 143.6, 95% interval 133–155. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$3.60cache read $0.18$18.001M
72.1
Good value (72.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
169.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.2
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 9 Nov 2026
149.0
From 171 benchmarks. Plenty of data behind it.
147 ±11
Coding IQ 146.6, 95% interval 135–158. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.20$40.00cache write $10.001M
73.1
Good value (73.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
44 ±9
Terse (44 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
173.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.2
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 9 Nov 2026
anthropic/claude-fable-5.2:batch · batch
149.0
From 171 benchmarks. Plenty of data behind it.
147 ±11
Coding IQ 146.6, 95% interval 135–158. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.10$20.00cache write $5.001M
73.1
Good value (73.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
44 ±9
Terse (44 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
173.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.2
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 9 Nov 2026
Inference by Anthropic
149.0
From 171 benchmarks. Plenty of data behind it.
147 ±11
Coding IQ 146.6, 95% interval 135–158. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.20$40.00cache write $10.001M
73.1
Good value (73.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
44 ±9
Terse (44 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
173.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.2
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 9 Nov 2026
anthropic/claude-fable-5.2:batch · batch
Inference by Anthropic
149.0
From 171 benchmarks. Plenty of data behind it.
147 ±11
Coding IQ 146.6, 95% interval 135–158. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.10$20.00cache write $5.001M
73.1
Good value (73.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
44 ±9
Terse (44 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
173.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.2
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 9 Nov 2026
fast
Inference by Anthropic
149.0
From 171 benchmarks. Plenty of data behind it.
147 ±11
Coding IQ 146.6, 95% interval 135–158. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$16.00cache read $0.40$80.00cache write $20.001M
73.1
Good value (73.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
44 ±9
Terse (44 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
173.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.2
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 9 Nov 2026
Inference by Microsoft Azure
149.0
From 171 benchmarks. Plenty of data behind it.
147 ±11
Coding IQ 146.6, 95% interval 135–158. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.20$40.00cache write $10.001M
73.1
Good value (73.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
44 ±9
Terse (44 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
173.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.2
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 9 Nov 2026
global
Inference by Google
149.0
From 171 benchmarks. Plenty of data behind it.
147 ±11
Coding IQ 146.6, 95% interval 135–158. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.20$40.00cache write $10.001M
73.1
Good value (73.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
44 ±9
Terse (44 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
173.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.40$40.00cache write $10.001.1M
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra:batch · batch
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001.1M
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra-pro:batch · batch+pro
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001.1M
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra-pro · pro
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.40$40.00cache write $10.001.1M
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
global
Inference by Amazon Web Services
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.40$40.00cache write $10.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
Inference by Microsoft Azure
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.40$40.00cache write $10.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra-pro · pro
Inference by Microsoft Azure
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.40$40.00cache write $10.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
Inference by OpenAI
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.40$40.00cache write $10.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra:batch · batch
Inference by OpenAI
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001.1M
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra-pro:batch · batch+pro
Inference by OpenAI
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001.1M
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
fast
Inference by OpenAI
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$16.00cache read $0.80$80.00cache write $20.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra-pro · fast+pro
Inference by OpenAI
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$16.00cache read $0.80$80.00cache write $20.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra-pro · pro
Inference by OpenAI
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.40$40.00cache write $10.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
openai/gpt-6.2-astra-pro · pro+ultrafast
Inference by OpenAI
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$24.00cache read $1.20$120.00cache write $30.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.2 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 17 Nov 2026
ultrafast
Inference by OpenAI
148.3
From 120 benchmarks. Plenty of data behind it.
142 ±9
Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$24.00cache read $1.20$120.00cache write $30.00922K
66.9
Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
172.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501.1M
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra:batch · batch
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251.1M
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra-pro:batch · batch+pro
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251.1M
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra-pro · pro
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501.1M
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
global
Inference by Amazon Web Services
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.50922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
Inference by Microsoft Azure
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.50922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra-pro · pro
Inference by Microsoft Azure
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.50922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.50922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra:batch · batch
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251.1M
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra-pro:batch · batch+pro
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251.1M
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
fast
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$20.00cache read $2.00$100.00cache write $25.00922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra-pro · fast+pro
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$20.00cache read $2.00$100.00cache write $25.00922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
flex
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.25922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra-pro · flex+pro
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.25922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra-pro · pro
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.50922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
openai/gpt-6-astra-pro · pro+ultrafast
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$60.00cache read $6.00$300.00cache write $75.00922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Astra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.
ultrafast
Inference by OpenAI
147.2
From 338 benchmarks. Plenty of data behind it.
139 ±6
Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$60.00cache read $6.00$300.00cache write $75.00922K
62.1
Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
37 ±8
Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
169.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
anthropic/claude-opus-5.5:batch · batch
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Amazon Web Services
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Anthropic
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
anthropic/claude-opus-5.5:batch · batch
Inference by Anthropic
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
fast
Inference by Anthropic
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.40$40.00cache write $10.001M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Microsoft Azure
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Claude Platform On AWS
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Google
146.6
From 329 benchmarks. Plenty of data behind it.
147 ±5
Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001M
66.8
Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
65 ±12
Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
168.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Hydrogen
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 12 Nov 2026
146.6
From 107 benchmarks. Plenty of data behind it.
126 ±11
Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.50cache read $0.075$7.501M
65.1
Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Hydrogen
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 12 Nov 2026
google/gemini-4-hydrogen:batch · batch
146.6
From 107 benchmarks. Plenty of data behind it.
126 ±11
Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.75cache read $0.0375$3.751M
65.1
Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Hydrogen
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 12 Nov 2026
Inference by Google
146.6
From 107 benchmarks. Plenty of data behind it.
126 ±11
Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.50cache read $0.075$7.501M
65.1
Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Hydrogen
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 12 Nov 2026
google/gemini-4-hydrogen:batch · batch+global
Inference by Google
146.6
From 107 benchmarks. Plenty of data behind it.
126 ±11
Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.75cache read $0.0375$3.751M
65.1
Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Hydrogen
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 12 Nov 2026
flex
Inference by Google
146.6
From 107 benchmarks. Plenty of data behind it.
126 ±11
Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.75cache read $0.0375$3.751M
65.1
Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Hydrogen
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 12 Nov 2026
global
Inference by Google
146.6
From 107 benchmarks. Plenty of data behind it.
126 ±11
Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.50cache read $0.075$7.501M
65.1
Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Hydrogen
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 12 Nov 2026
global/priority
Inference by Google
146.6
From 107 benchmarks. Plenty of data behind it.
126 ±11
Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.70cache read $0.135$13.501M
65.1
Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Hydrogen
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
New
Released 12 Nov 2026
priority
Inference by Google
146.6
From 107 benchmarks. Plenty of data behind it.
126 ±11
Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.70cache read $0.135$13.501M
65.1
Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501.1M
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol:batch · batch
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol-pro:batch · batch+pro
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol-pro · pro
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501.1M
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Amazon Web Services
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Microsoft Azure
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol-pro · pro
Inference by Microsoft Azure
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol:batch · batch
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol-pro:batch · batch+pro
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
fast
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.00922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol-pro · fast+pro
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.00922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
flex
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.25922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol-pro · flex+pro
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.25922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol-pro · pro
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-sol-pro · pro+ultrafast
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$12.00cache read $0.60$60.00cache write $15.00922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
ultrafast
Inference by OpenAI
145.6
From 239 benchmarks. Plenty of data behind it.
135 ±16
Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$12.00cache read $0.60$60.00cache write $15.00922K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
51 ±11
Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
168.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Argon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
145.4
From 217 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.001M
66.0
Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
66 ±9
Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate.
166.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Argon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
google/gemini-4-argon:batch · batch
145.4
From 217 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.001M
66.0
Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
66 ±9
Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate.
166.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Argon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Google
145.4
From 217 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.001M
66.0
Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
66 ±9
Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate.
166.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Argon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
google/gemini-4-argon:batch · batch+global
Inference by Google
145.4
From 217 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.001M
66.0
Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
66 ±9
Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate.
166.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Argon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
flex
Inference by Google
145.4
From 217 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.001M
66.0
Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
66 ±9
Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate.
166.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Argon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Google
145.4
From 217 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.001M
66.0
Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
66 ±9
Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate.
166.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Argon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global/priority
Inference by Google
145.4
From 217 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$3.60cache read $0.18$18.001M
66.0
Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
66 ±9
Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate.
166.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 4 Argon
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
priority
Inference by Google
145.4
From 217 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$3.60cache read $0.18$18.001M
66.0
Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
66 ±9
Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate.
166.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
145.3
From 253 benchmarks. Plenty of data behind it.
138 ±5
Coding IQ 137.9, 95% interval 133–143. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $0.25$50.00cache write $12.501M
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
63 ±9
Chatty (63 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
167.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
anthropic/claude-fable-5.1:batch · batch
145.3
From 253 benchmarks. Plenty of data behind it.
138 ±5
Coding IQ 137.9, 95% interval 133–143. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.125$25.00cache write $6.251M
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
63 ±9
Chatty (63 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
167.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
Inference by Anthropic
145.3
From 253 benchmarks. Plenty of data behind it.
138 ±5
Coding IQ 137.9, 95% interval 133–143. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $0.25$50.00cache write $12.501M
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
63 ±9
Chatty (63 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
167.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
anthropic/claude-fable-5.1:batch · batch
Inference by Anthropic
145.3
From 253 benchmarks. Plenty of data behind it.
138 ±5
Coding IQ 137.9, 95% interval 133–143. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.125$25.00cache write $6.251M
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
63 ±9
Chatty (63 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
167.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
Inference by Microsoft Azure
145.3
From 253 benchmarks. Plenty of data behind it.
138 ±5
Coding IQ 137.9, 95% interval 133–143. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $0.25$50.00cache write $12.501M
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
63 ±9
Chatty (63 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
167.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
global
Inference by Google
145.3
From 253 benchmarks. Plenty of data behind it.
138 ±5
Coding IQ 137.9, 95% interval 133–143. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $0.25$50.00cache write $12.501M
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
63 ±9
Chatty (63 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
167.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen4 Max
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
New
Released 11 Nov 2026
144.9
From 40 benchmarks. Plenty of data behind it.
≈ 124 ±12
Coding IQ 124.3, 95% interval 112–137. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.425cache read $0.0713$4.275cache write $1.78131M
61.7
Good value (61.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measuredNot measured
Qwen4 Max
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
New
Released 11 Nov 2026
Inference by Alibaba Cloud
144.9
From 40 benchmarks. Plenty of data behind it.
≈ 124 ±12
Coding IQ 124.3, 95% interval 112–137. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.425cache read $0.0713$4.275cache write $1.78131M
61.7
Good value (61.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measuredNot measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501.1M
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol:batch · batch
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol-pro:batch · batch+pro
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol-pro · pro
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501.1M
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Amazon Web Services
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Microsoft Azure
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol-pro · pro
Inference by Microsoft Azure
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by OpenAI
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol:batch · batch
Inference by OpenAI
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol-pro:batch · batch+pro
Inference by OpenAI
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
fast
Inference by OpenAI
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.00922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol-pro · fast+pro
Inference by OpenAI
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.00922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol-pro · pro
Inference by OpenAI
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.2-sol-pro · pro+ultrafast
Inference by OpenAI
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$6.00cache read $0.30$30.00cache write $7.50922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
GPT-6.2 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
ultrafast
Inference by OpenAI
144.3
From 104 benchmarks. Plenty of data behind it.
139 ±9
Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$6.00cache read $0.30$30.00cache write $7.50922K
69.3
Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
Not measured
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
anthropic/claude-opus-5-fast · fast
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
anthropic/claude-opus-5:batch · batch
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.50cache read $0.25$12.50cache write $3.1251M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
Inference by Amazon Web Services
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
Inference by Anthropic
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
anthropic/claude-opus-5:batch · batch
Inference by Anthropic
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.50cache read $0.25$12.50cache write $3.1251M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
fast
Inference by Anthropic
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
global
Inference by Microsoft Azure
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
Inference by Claude Platform On AWS
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
global
Inference by Google
143.9
From 389 benchmarks. Plenty of data behind it.
134 ±6
Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
64.6
Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it.
61 ±9
Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
164.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Sonnet 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
143.9
From 270 benchmarks. Plenty of data behind it.
140 ±5
Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501M
66.7
Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
38 ±9
Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Sonnet 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
anthropic/claude-sonnet-5.5:batch · batch
143.9
From 270 benchmarks. Plenty of data behind it.
140 ±5
Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251M
66.7
Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
38 ±9
Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Sonnet 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Amazon Web Services
143.9
From 270 benchmarks. Plenty of data behind it.
140 ±5
Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501M
66.7
Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
38 ±9
Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Sonnet 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Anthropic
143.9
From 270 benchmarks. Plenty of data behind it.
140 ±5
Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501M
66.7
Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
38 ±9
Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Sonnet 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
anthropic/claude-sonnet-5.5:batch · batch
Inference by Anthropic
143.9
From 270 benchmarks. Plenty of data behind it.
140 ±5
Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251M
66.7
Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
38 ±9
Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Sonnet 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Microsoft Azure
143.9
From 270 benchmarks. Plenty of data behind it.
140 ±5
Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501M
66.7
Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
38 ±9
Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Sonnet 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Claude Platform On AWS
143.9
From 270 benchmarks. Plenty of data behind it.
140 ±5
Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501M
66.7
Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
38 ±9
Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Sonnet 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Google
143.9
From 270 benchmarks. Plenty of data behind it.
140 ±5
Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.501M
66.7
Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
38 ±9
Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
166.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
143.0
From 383 benchmarks. Plenty of data behind it.
132 ±5
Coding IQ 132.1, 95% interval 128–137. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501M
57.1
Good value (57.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±10
Typical (53 ±10). Generates about as many tokens as models with similar coding results. Measured at default effort. A fair amount of data behind it.
164.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
anthropic/claude-fable-5:batch · batch
143.0
From 383 benchmarks. Plenty of data behind it.
132 ±5
Coding IQ 132.1, 95% interval 128–137. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
57.1
Good value (57.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±10
Typical (53 ±10). Generates about as many tokens as models with similar coding results. Measured at default effort. A fair amount of data behind it.
164.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
Inference by Anthropic
143.0
From 383 benchmarks. Plenty of data behind it.
132 ±5
Coding IQ 132.1, 95% interval 128–137. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501M
57.1
Good value (57.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±10
Typical (53 ±10). Generates about as many tokens as models with similar coding results. Measured at default effort. A fair amount of data behind it.
164.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
Inference by Microsoft Azure
143.0
From 383 benchmarks. Plenty of data behind it.
132 ±5
Coding IQ 132.1, 95% interval 128–137. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501M
57.1
Good value (57.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±10
Typical (53 ±10). Generates about as many tokens as models with similar coding results. Measured at default effort. A fair amount of data behind it.
164.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
Inference by Claude Platform On AWS
143.0
From 383 benchmarks. Plenty of data behind it.
132 ±5
Coding IQ 132.1, 95% interval 128–137. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501M
57.1
Good value (57.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±10
Typical (53 ±10). Generates about as many tokens as models with similar coding results. Measured at default effort. A fair amount of data behind it.
164.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Fable 5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.
global
Inference by Google
143.0
From 383 benchmarks. Plenty of data behind it.
132 ±5
Coding IQ 132.1, 95% interval 128–137. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501M
57.1
Good value (57.1). Cheaper per task than its benchmark results predict. Plenty of data behind it.
53 ±10
Typical (53 ±10). Generates about as many tokens as models with similar coding results. Measured at default effort. A fair amount of data behind it.
164.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Muse Wonder
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
142.8
From 123 benchmarks. Plenty of data behind it.
134 ±8
Coding IQ 134.3, 95% interval 127–142. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.001M
54.0
Good value (54.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
54 ±10
Typical (54 ±10). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
161.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Muse Wonder
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Google
142.8
From 123 benchmarks. Plenty of data behind it.
134 ±8
Coding IQ 134.3, 95% interval 127–142. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.001M
54.0
Good value (54.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
54 ±10
Typical (54 ±10). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
161.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Muse Wonder
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Meta
142.8
From 123 benchmarks. Plenty of data behind it.
134 ±8
Coding IQ 134.3, 95% interval 127–142. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.001M
54.0
Good value (54.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
54 ±10
Typical (54 ±10). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
161.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.253cache read $0.18$8.801M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
moonshotai/kimi-k3.1-pro:batch · batch
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.372cache read $0.10$5.5831M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Akash
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.75cache read $0.18$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Alibaba Cloud
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
prime
Inference by Alibaba Cloud
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us-east-1
Inference by Amazon Web Services
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.40cache read $0.22$22.00cache write $5.501M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us-east-2
Inference by Amazon Web Services
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.00cache write $5.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
mxfp4
Inference by Decart
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.20cache read $0.12$6.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
bf16
Inference by DeepInfra
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.90cache read $0.19$9.501M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Digitalocean
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.90cache read $0.19$9.501M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Fireworks AI
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fast · mxfp4
Inference by Inferencenet
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.70cache read $0.27$13.501M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
mxfp4
Inference by Inferencenet
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.80cache read $0.198$9.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
mxfp4/us · mxfp4
Inference by Inferencenet
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.80cache read $0.198$9.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us/fast · mxfp4
Inference by Inferencenet
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.70cache read $0.27$13.501M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Makora
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$9.491M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
mxfp4
Inference by Modal
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
highspeed · mxfp4
Inference by Moonshot AI
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.20$20.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
mxfp4
Inference by Moonshot AI
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
moonshotai/kimi-k3.1-pro:batch · batch+fp8 · fp8
Inference by Morph
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.372cache read $0.10$5.5831M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Morph
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.375cache read $0.172$9.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Nebius AI Studio
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Parasail
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
mxfp4
Inference by PrimaLabs
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Relace
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2403cache read $0.20$8.40231M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by River
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
moonshotai/kimi-k3.1-pro:batch · batch+fp4 · fp4
Inference by Sail Research
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.70cache read $0.10$6.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Sail Research
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.4025cache read $0.187$9.351M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4/us · fp4
Inference by Sail Research
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.0425cache read $0.187$9.351M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Together AI
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
mxfp4
Inference by Wafer
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.80cache read $0.18$9.001M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K3.1 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us · mxfp4
Inference by Wafer
142.7
From 161 benchmarks. Plenty of data behind it.
131 ±10
Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.253cache read $0.18$8.801M
49.2
Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
51 ±9
Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
161.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra:batch · batch
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.50cache read $0.025$2.50cache write $0.6251.1M
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra-pro:batch · batch+pro
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.50cache read $0.025$2.50cache write $0.6251.1M
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra-pro · pro
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.251.1M
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Microsoft Azure
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.25922K
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra-pro · pro
Inference by Microsoft Azure
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.25922K
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by OpenAI
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.25922K
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra:batch · batch
Inference by OpenAI
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.50cache read $0.025$2.50cache write $0.6251.1M
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra-pro:batch · batch+pro
Inference by OpenAI
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.50cache read $0.025$2.50cache write $0.6251.1M
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
fast
Inference by OpenAI
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra-pro · fast+pro
Inference by OpenAI
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.10$10.00cache write $2.50922K
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
flex
Inference by OpenAI
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.50cache read $0.025$2.50cache write $0.625922K
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra-pro · flex+pro
Inference by OpenAI
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.50cache read $0.025$2.50cache write $0.625922K
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6.1 Terra
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6.1-terra-pro · pro
Inference by OpenAI
142.5
From 130 benchmarks. Plenty of data behind it.
130 ±13
Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.05$5.00cache write $1.25922K
54.7
Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it.
52 ±9
Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
167.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.501.1M
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol:batch · batch
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.10$5.00cache write $1.251.1M
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol-pro:batch · batch+pro
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.10$5.00cache write $1.251.1M
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol-pro · pro
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.501.1M
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Amazon Web Services
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.50922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Microsoft Azure
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.50922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol-pro · pro
Inference by Microsoft Azure
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.50922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by OpenAI
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.50922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol:batch · batch
Inference by OpenAI
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.10$5.00cache write $1.251.1M
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol-pro:batch · batch+pro
Inference by OpenAI
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.10$5.00cache write $1.251.1M
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
fast
Inference by OpenAI
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.00922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol-pro · fast+pro
Inference by OpenAI
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.00922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
flex
Inference by OpenAI
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.10$5.00cache write $1.25922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol-pro · flex+pro
Inference by OpenAI
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.10$5.00cache write $1.25922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-6-sol-pro · pro
Inference by OpenAI
142.5
From 210 benchmarks. Plenty of data behind it.
130 ±7
Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.50922K
63.4
Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
163.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.001.1M
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol:batch · batch
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.501.1M
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol-pro:batch · batch+pro
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.501.1M
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol-pro · pro
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.001.1M
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
global
Inference by Amazon Web Services
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.00922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
Inference by Microsoft Azure
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.00922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol-pro · pro
Inference by Microsoft Azure
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.00922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
Inference by OpenAI
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.00922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol:batch · batch
Inference by OpenAI
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.501.1M
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol-pro:batch · batch+pro
Inference by OpenAI
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.501.1M
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
fast
Inference by OpenAI
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.80$40.00cache write $10.00922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol-pro · fast+pro
Inference by OpenAI
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$8.00cache read $0.80$40.00cache write $10.00922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
flex
Inference by OpenAI
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.50922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol-pro · flex+pro
Inference by OpenAI
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $0.20$10.00cache write $2.50922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.6 Sol
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.
openai/gpt-5.6-sol-pro · pro
Inference by OpenAI
141.6
From 504 benchmarks. Plenty of data behind it.
125 ±5
Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00cache read $0.40$20.00cache write $5.00922K
59.9
Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
41 ±8
Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it.
163
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.5 Pro
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
140.6
From 26 benchmarks. Plenty of data behind it.
Not enough data yet$30.00$180.001.1M
48.5
Fair value (48.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
Not measured
162.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.5 Pro
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-5.5-pro:batch · batch
140.6
From 26 benchmarks. Plenty of data behind it.
Not enough data yet$15.00$90.001.1M
48.5
Fair value (48.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
Not measured
162.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.5 Pro
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by OpenAI
140.6
From 26 benchmarks. Plenty of data behind it.
Not enough data yet$30.00$180.00922K
48.5
Fair value (48.5). Costs per task about what its benchmark results predict. A fair amount of data behind it.
Not measured
162.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Muse Spark 1.3
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
140.3
From 224 benchmarks. Plenty of data behind it.
125 ±6
Coding IQ 125.1, 95% interval 119–132. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.25cache read $0.15$4.251M
52.1
Fair value (52.1). Costs per task about what its benchmark results predict. Plenty of data behind it.
59 ±9
Chatty (59 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
158.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Muse Spark 1.3
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
global
Inference by Google
140.3
From 224 benchmarks. Plenty of data behind it.
125 ±6
Coding IQ 125.1, 95% interval 119–132. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.25cache read $0.15$4.251M
52.1
Fair value (52.1). Costs per task about what its benchmark results predict. Plenty of data behind it.
59 ±9
Chatty (59 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
158.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Muse Spark 1.3
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Inference by Meta
140.3
From 224 benchmarks. Plenty of data behind it.
125 ±6
Coding IQ 125.1, 95% interval 119–132. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.25cache read $0.15$4.251M
52.1
Fair value (52.1). Costs per task about what its benchmark results predict. Plenty of data behind it.
59 ±9
Chatty (59 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate.
158.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT 6We compared the version names in our catalogue.
139.8
From 476 benchmarks. Plenty of data behind it.
122 ±12
Coding IQ 121.5, 95% interval 110–134. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$30.001.1M
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
28 ±8
Terse (28 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
160.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT 6We compared the version names in our catalogue.
openai/gpt-5.5:batch · batch
139.8
From 476 benchmarks. Plenty of data behind it.
122 ±12
Coding IQ 121.5, 95% interval 110–134. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.50cache read $0.25$15.001.1M
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
28 ±8
Terse (28 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
160.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT 6We compared the version names in our catalogue.
Inference by Microsoft Azure
139.8
From 476 benchmarks. Plenty of data behind it.
122 ±12
Coding IQ 121.5, 95% interval 110–134. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$30.00922K
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
28 ±8
Terse (28 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
160.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT 6We compared the version names in our catalogue.
Inference by OpenAI
139.8
From 476 benchmarks. Plenty of data behind it.
122 ±12
Coding IQ 121.5, 95% interval 110–134. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$30.00922K
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
28 ±8
Terse (28 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
160.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT 6We compared the version names in our catalogue.
fast
Inference by OpenAI
139.8
From 476 benchmarks. Plenty of data behind it.
122 ±12
Coding IQ 121.5, 95% interval 110–134. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$12.50cache read $1.25$75.00922K
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
28 ±8
Terse (28 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
160.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 5.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGPT 6We compared the version names in our catalogue.
flex
Inference by OpenAI
139.8
From 476 benchmarks. Plenty of data behind it.
122 ±12
Coding IQ 121.5, 95% interval 110–134. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.50cache read $0.25$15.00922K
64.0
Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it.
28 ±8
Terse (28 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
160.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
anthropic/claude-opus-4.8-fast · fast
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.50Not published by the provider
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
anthropic/claude-opus-4.8:batch · batch
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.50cache read $0.25$12.50cache write $3.1251M
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
Inference by Amazon Web Services
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
Inference by Anthropic
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
fast
Inference by Anthropic
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$10.00cache read $1.00$50.00cache write $12.501M
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
global
Inference by Microsoft Azure
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
Inference by Claude Platform On AWS
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.8
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.
global
Inference by Google
139.7
From 478 benchmarks. Plenty of data behind it.
121 ±4
Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$5.00cache read $0.50$25.00cache write $6.251M
52.9
Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
40 ±10
Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
159.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.15cache read $0.003$0.601M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
deepseek/deepseek-v4.1-flash-1017:batch · batch
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.035cache read $0.0029$0.1491M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Alibaba Cloud
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.18cache read $0.0036$0.721M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Ambient
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.045cache read $0.0025$0.251M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Ant Digital
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.12cache read $0.0024$0.481M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Baidu
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0977cache read $0.0021$0.39441M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Baseten
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.007$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fast
Inference by Baseten
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60cache read $0.014$2.401M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Cloudflare Workers AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.10cache read $0.003$0.601M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Cohere
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.115cache read $0.0025$0.401M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Coreweave
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.006$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Crusoe
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.29cache read $0.007$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by DeepInfra
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.098cache read $0.0029$0.2941M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8/flex · fp8
Inference by DeepInfra
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0784cache read $0.0024$0.23521M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8/priority · fp8
Inference by DeepInfra
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147cache read $0.0044$0.4411M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by DeepSeek
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.15cache read $0.003$0.601M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Dekallm
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.10cache read $0.003$0.601M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Digitalocean
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.15cache read $0.003$0.601M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Fireworks AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.006$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us
Inference by Fireworks AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.45cache read $0.009$1.801M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Inceptron
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.10cache read $0.006$0.601M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Inferencenet
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.08cache read $0.008$0.6661M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us · fp8
Inference by Inferencenet
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.08cache read $0.008$0.6661M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Ionstream
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.21cache read $0.0105$0.621M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Makora
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.15cache read $0.006$0.801M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Mistral AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.007$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Modal
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.25cache read $0.005$1.001M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Modelrun
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.17cache read $0.006$0.671M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
deepseek/deepseek-v4.1-flash-1017:batch · batch+fp8 · fp8
Inference by Morph
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.036cache read $0.003$0.151M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fast · fp8
Inference by Morph
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60cache read $0.014$2.401M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Morph
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.059cache read $0.005$0.521M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Near AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.006$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Nebius AI Studio
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.18cache read $0.0036$0.721M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Novita AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.165cache read $0.0033$0.661M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Openinference
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.05cache read $0.002$0.751M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Parasail
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.006$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us
Inference by Particle AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.06cache read $0.03$0.551M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us/flex
Inference by Particle AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.03cache read $0.015$0.2751M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Pearl Research
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.05cache read $0.003$0.301M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by PrimaLabs
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.24cache read $0.0048$0.961M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Primeintellect
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0568cache read $0.0031$0.701M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Reka AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.006$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Relace
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.12cache read $0.003$0.501M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by River
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.12cache read $0.0024$0.481M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
deepseek/deepseek-v4.1-flash-1017:batch · batch+fp4 · fp4
Inference by Sail Research
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.05cache read $0.004$0.151M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4
Inference by Sail Research
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.04cache read $0.007$0.101M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp4/us · fp4
Inference by Sail Research
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.04cache read $0.007$0.101M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by SiliconFlow
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.15cache read $0.003$0.601M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by StepFun
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105cache read $0.0021$0.421M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Streamlake
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.075cache read $0.0015$0.301M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Inference by Together AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.006$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
deepseek/deepseek-v4.1-flash-1017:batch · batch
Inference by Together AI
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.035cache read $0.0029$0.1491M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
fp8
Inference by Venice
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.0075$1.201M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
mxfp4
Inference by Wafer
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.066cache read $0.0065$0.501M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V4.1 Flash 1017
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
us · mxfp4
Inference by Wafer
139.7
From 97 benchmarks. Plenty of data behind it.
127 ±8
Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.036cache read $0.01$0.501M
71.4
Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
55 ±9
Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
159.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence)

How this data works