Skip to content

Listings

All listings: NanoGPT

Every listing for the models on page 9 of 40 of NanoGPT, including each model's other versions and routes. Back to NanoGPT

Sold by NanoGPT

73 listings on this page sold by NanoGPT.

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
Mistral Medium 3.5
🇫🇷
Maker headquartered in FR, according to the company's own website. The citation is on the model page.
115.3
From 173 benchmarks. Plenty of data behind it.
99 ±15
Coding IQ 98.6, 95% interval 84–114. 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 $0.75$7.50256K
12.0
Poor value (12.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
141.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Mistral Medium 3.5
🇫🇷
Maker headquartered in FR, according to the company's own website. The citation is on the model page.
mistralai/mistral-medium-3.5:thinking · thinking
115.3
From 173 benchmarks. Plenty of data behind it.
99 ±15
Coding IQ 98.6, 95% interval 84–114. 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 $0.75$7.50256K
12.0
Poor value (12.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
141.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 2.5 Pro TTS
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
gemini-2.5-pro-preview-tts
≈ 115
From 16 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Reka Flash 3
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
115.0
From 12 benchmarks. A fair amount of data behind it.
Not enough data yet$0.105$0.2166K
38.8
Poor value (38.8). Costs more per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Reka Flash 3
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
UNK
Inference by Reka AI
115.0
From 12 benchmarks. A fair amount of data behind it.
Not enough data yet$0.105$0.21Not published by the provider
38.8
Poor value (38.8). Costs more per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Qwen3 30B A3B Instruct
🇨🇳
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.
qwen/qwen3-30b-a3b-instruct-2507
114.9
From 100 benchmarks. Plenty of data behind it.
≈ 86 ±20
Coding IQ 86.4, 95% interval 67–106. 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.
$0.20cache read $0.10$0.50256K
36.5
Poor value (36.5). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measured
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Newer version
Newer versionDeepSeek V4We compared the version names in our catalogue.
deepseek-v3-0324 · 0324
114.5
From 492 benchmarks. Plenty of data behind it.
85 ±7
Coding IQ 85.3, 95% interval 79–92. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.29$1.16128K
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
47 ±9
Typical (47 ±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.
135.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Newer version
Newer versionDeepSeek V4We compared the version names in our catalogue.
deepseek-v3-0324 · 0324 · fp8 · 🇸🇬 SG
Inference by SiliconFlow
114.5
From 492 benchmarks. Plenty of data behind it.
85 ±7
Coding IQ 85.3, 95% interval 79–92. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2625$1.05Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
47 ±9
Typical (47 ±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.
135.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Newer version
Newer versionDeepSeek V4We compared the version names in our catalogue.
deepseek-v3-0324 · 0324 · bf16 · GLB
Inference by Uomi
114.5
From 492 benchmarks. Plenty of data behind it.
85 ±7
Coding IQ 85.3, 95% interval 79–92. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2756$1.1025Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
47 ±9
Typical (47 ±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.
135.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3 30B A3B
🇨🇳
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.
114.4
From 142 benchmarks. Plenty of data behind it.
90 ±10
Coding IQ 90.3, 95% interval 80–101. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.12$0.5041K
35.8
Poor value (35.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
138.3
Epoch Capabilities IndexSetting: thinkingSource: Epoch AI Benchmarking Hub API
Seedream 5.0 Probytedance/seedream-v5.0-pro
≈ 114
From 38 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Seedream 5.0 Probytedance/seedream-v5.0-pro/edit · edit
≈ 114
From 38 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Seedream 5.0 Probytedance/seedream/v5/pro/text-to-image · text-to-image
≈ 114
From 38 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Qwen3 14B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
114.1
From 210 benchmarks. Plenty of data behind it.
85 ±9
Coding IQ 84.9, 95% interval 76–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.08cache read $0.04$0.2441K
33.7
Poor value (33.7). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measured
136.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3 14B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
fp8 · 🇺🇸 US
Inference by DeepInfra
114.1
From 210 benchmarks. Plenty of data behind it.
85 ±9
Coding IQ 84.9, 95% interval 76–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.126$0.252Not published by the provider
33.7
Poor value (33.7). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measured
136.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3 14B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
fp8 · 🇪🇺 EU
Inference by Greenference
114.1
From 210 benchmarks. Plenty of data behind it.
85 ±9
Coding IQ 84.9, 95% interval 76–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0998cache read $0.0499$0.2993Not published by the provider
33.7
Poor value (33.7). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measured
136.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3 14B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
int4 · 🇪🇺 EU
Inference by Nextbit
114.1
From 210 benchmarks. Plenty of data behind it.
85 ±9
Coding IQ 84.9, 95% interval 76–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105$0.231Not published by the provider
33.7
Poor value (33.7). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measured
136.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 4o
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
113.7
From 761 benchmarks. Plenty of data behind it.
77 ±10
Coding IQ 77.4, 95% interval 67–87. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.50cache read $1.25$10.00128K
24.6
Poor value (24.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
128.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 4o
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-4o-2024-08-06 · 2024-08-06
113.7
From 761 benchmarks. Plenty of data behind it.
77 ±10
Coding IQ 77.4, 95% interval 67–87. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.50cache read $1.25$10.00128K
24.6
Poor value (24.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
128.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT 4o
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/gpt-4o-2024-11-20 · 2024-11-20
113.7
From 761 benchmarks. Plenty of data behind it.
77 ±10
Coding IQ 77.4, 95% interval 67–87. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.50cache read $1.25$10.00128K
24.6
Poor value (24.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
128.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemini 2.5 Flash-Lite
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionGemini 3.5 Flash LiteWe compared the version names in our catalogue.
113.6
From 170 benchmarks. Plenty of data behind it.
≈ 80 ±12
Coding IQ 79.7, 95% interval 68–92. 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.
$0.10cache read $0.01$0.401M
42.3
Poor value (42.3). Costs more 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.
133
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Mai Image 2.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionMAI Image 2.7We compared the version names in our catalogue.
microsoft/mai-image-2.5/edit · edit
113.6
From 15 benchmarks. A fair amount of data behind it.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Mai Image 2.5
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
Newer version
Newer versionMAI Image 2.7We compared the version names in our catalogue.
microsoft/mai-image-2.5/text-to-image · text-to-image
113.6
From 15 benchmarks. A fair amount of data behind it.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
LFM2.5 2.6B
🇺🇸
Maker headquartered in US, 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.
113.3
From 9 benchmarks. A fair amount of data behind it.
Not enough data yet$0.10cache read $0.05$0.208KNot enough data yetNot measuredNot measured
LFM2.5 2.6B
🇺🇸
Maker headquartered in US, 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.
GLB
Inference by Kitani
113.3
From 9 benchmarks. A fair amount of data behind it.
Not enough data yet$0.0945cache read $0.0473$0.189Not published by the providerNot enough data yetNot measuredNot measured
Mistral Small 3.2 24B Instruct
🇫🇷
Maker headquartered in FR, 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.
113.3
From 50 benchmarks. Plenty of data behind it.
Not enough data yet$0.075$0.20128K
58.5
Good value (58.5). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Mistral Small 3.2 24B Instruct
🇫🇷
Maker headquartered in FR, 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.
int4 · GLB
Inference by Kitani
113.3
From 50 benchmarks. Plenty of data behind it.
Not enough data yet$0.189cache read $0.0945$0.399Not published by the provider
58.5
Good value (58.5). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Qwen3.5 9B
🇨🇳
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.
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.10cache read $0.05$0.15262K
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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 · GLB
Inference by Darkbloom
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.084$0.1365Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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 · 🇺🇸 US
Inference by DeepInfra
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105$0.1575Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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 · 🇪🇺 EU
Inference by Greenference
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.042cache read $0.021$0.0683Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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.
🇪🇺 EU
Inference by Ionos
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1155$0.1785Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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 · GLB
Inference by Mixlayer
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105$0.42Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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 · 🇺🇸 US
Inference by Parasail
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105$0.2625Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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 · UNK
Inference by Primeintellect
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1785cache read $0.1607$0.2625Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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 · 🇸🇬 SG
Inference by SiliconFlow
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105$0.1575Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 9B
🇨🇳
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 Together AI
113.2
From 223 benchmarks. Plenty of data behind it.
86 ±8
Coding IQ 85.6, 95% interval 77–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1785$0.2625Not published by the provider
55.4
Good value (55.4). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
139.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen Image 3.0 Pro
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
qwen-image-3-pro
≈ 113
From 32 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Devstral Small 2505
🇫🇷
Maker headquartered in FR, according to the company's own website. The citation is on the model page.
2505
112.7
From 14 benchmarks. A fair amount of data behind it.
≈ 83 ±10stale
Coding IQ 82.9, 95% interval 73–93. 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. Its newest result is over a year old.
$0.06cache read $0.03$0.0633K
46.0
Poor value (46.0). Costs more per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Nova 2 Lite
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
amazon/nova-2-lite-v1
112.6
From 78 benchmarks. Plenty of data behind it.
94 ±41
Coding IQ 94.4, 95% interval 53–136. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.51cache read $0.255$4.251M
38.2
Poor value (38.2). Costs more per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Qwen3 8B
🇨🇳
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.
112.4
From 182 benchmarks. Plenty of data behind it.
80 ±12
Coding IQ 80.3, 95% interval 68–92. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.47cache read $0.235$0.4741K
58.0
Good value (58.0). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measured
132.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3 8B
🇨🇳
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 · 🇪🇺 EU
Inference by Greenference
112.4
From 182 benchmarks. Plenty of data behind it.
80 ±12
Coding IQ 80.3, 95% interval 68–92. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.084cache read $0.042$0.2625Not published by the provider
58.0
Good value (58.0). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measured
132.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Grok Imagine Video
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
≈ 112
From 7 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Grok Imagine Video
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
grok-imagine-video-reference-to-video · reference-to-video
≈ 112
From 7 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Llama 4 Maverick
🇺🇸
Maker headquartered in US, 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.
111.8
From 350 benchmarks. Plenty of data behind it.
78 ±9
Coding IQ 78.1, 95% interval 69–87. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.35cache read $0.17$1.001M
42.1
Poor value (42.1). Costs more 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.
132.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Llama 4 Maverick
🇺🇸
Maker headquartered in US, 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 · 🇺🇸 US
Inference by Parasail
111.8
From 350 benchmarks. Plenty of data behind it.
78 ±9
Coding IQ 78.1, 95% interval 69–87. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.3675cache read $0.1785$1.05Not published by the provider
42.1
Poor value (42.1). Costs more 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.
132.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Llama 4 Maverick
🇺🇸
Maker headquartered in US, 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 · GLB
Inference by Scx
111.8
From 350 benchmarks. Plenty of data behind it.
78 ±9
Coding IQ 78.1, 95% interval 69–87. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.5565$1.701Not published by the provider
42.1
Poor value (42.1). Costs more 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.
132.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen Max
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
111.6
From 56 benchmarks. Plenty of data behind it.
≈ 82 ±14
Coding IQ 81.7, 95% interval 68–95. 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.
$1.5997cache read $0.7999$6.39232K
42.6
Poor value (42.6). Costs more per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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.
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.018cache read $0.009$0.09131K
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · GLB
Inference by Akash
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0189cache read $0.0095$0.0945Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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
Inference by Coreweave
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0315$0.1365Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · GLB
Inference by Darkbloom
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0179$0.0893Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · 🇺🇸 US
Inference by DeepInfra
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0315cache read $0.0158$0.147Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · 🇮🇩 ID
Inference by Dekallm
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0305$0.147Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · 🇪🇺 EU
Inference by Greenference
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0095cache read $0.0047$0.0473Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 Groq
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0788cache read $0.0394$0.315Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · 🇦🇪 AE
Inference by Hyperfusion
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.021$0.105Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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.
UNK
Inference by Ollama
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0735cache read $0.0368$0.315Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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
Inference by Parasail
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0315cache read $0.021$0.1575Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · UNK
Inference by Primeintellect
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.21cache read $0.189$0.315Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · 🇸🇬 SG
Inference by SiliconFlow
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.042$0.189Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT OSS 20b
🇺🇸
Maker headquartered in US, 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 · GLB
Inference by Uomi
111.3
From 251 benchmarks. Plenty of data behind it.
88 ±12
Coding IQ 88.2, 95% interval 76–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.032$0.1488Not published by the provider
48.9
Fair value (48.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
40 ±9
Terse (40 ±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.
134.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Granite 4.2 8B
🇺🇸
Maker headquartered in US, 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.
110.9
From 57 benchmarks. Plenty of data behind it.
90 ±17
Coding IQ 90.2, 95% interval 73–108. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.06cache read $0.015$0.15131KNot enough data yetNot measuredNot measured
Granite 4.2 8B
🇺🇸
Maker headquartered in US, 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 · 🇺🇸 US
Inference by Coreweave
110.9
From 57 benchmarks. Plenty of data behind it.
90 ±17
Coding IQ 90.2, 95% interval 73–108. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.063cache read $0.0158$0.1575Not published by the providerNot enough data yetNot measuredNot measured
Granite 4.2 8B
🇺🇸
Maker headquartered in US, 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 · 🇺🇸 US
Inference by DeepInfra
110.9
From 57 benchmarks. Plenty of data behind it.
90 ±17
Coding IQ 90.2, 95% interval 73–108. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.063cache read $0.0158$0.2625Not published by the providerNot enough data yetNot measuredNot measured
Granite 4.2 8B
🇺🇸
Maker headquartered in US, 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.
GLB
Inference by Kitani
110.9
From 57 benchmarks. Plenty of data behind it.
90 ±17
Coding IQ 90.2, 95% interval 73–108. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0945cache read $0.0473$0.1418Not published by the providerNot enough data yetNot measuredNot measured
Muse Image
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
meta/muse-image/edit · edit
≈ 111
From 15 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Muse Image
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
meta/muse-image/text-to-image · text-to-image
≈ 111
From 15 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Veo 3.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
veo3-1-video · audio+image
≈ 111
From 6 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Veo 3.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
veo3-1-extend · extend
≈ 111
From 6 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Veo 3.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
veo3-1-video · image+no-audio
≈ 111
From 6 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Veo 3.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
veo3-1-video · text+1080p
≈ 111
From 6 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured
Veo 3.1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
veo3-1-video · text+720p
≈ 111
From 6 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet——Not published by the providerNot enough data yetNot measuredNot measured

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

How this data works