Skip to content

Listings

All listings: NanoGPT

Every listing for the models on page 5 of 40 of NanoGPT, including each model's other versions and routes. Back to NanoGPT

Sold by NanoGPT

206 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
Doubao Seed 2.1 Turbobytedance-seed/seed-2-1-turbo
128.5
From 60 benchmarks. Plenty of data behind it.
Not enough data yet$0.50cache read $0.25$2.50262KNot enough data yetNot measuredNot measured
Kimi K2.5
🇨🇳
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.
Newer version
Newer versionKimi K3.1We compared the version names in our catalogue.
128.5
From 401 benchmarks. Plenty of data behind it.
103 ±5
Coding IQ 102.9, 95% interval 98–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.557cache read $0.0912$2.832262K
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±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.
148.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.5
🇨🇳
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.
Newer version
Newer versionKimi K3.1We compared the version names in our catalogue.
moonshotai/kimi-k2.5:thinking · thinking
128.5
From 401 benchmarks. Plenty of data behind it.
103 ±5
Coding IQ 102.9, 95% interval 98–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.6265cache read $0.1053$3.2187262K
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±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.
148.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.5
🇨🇳
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.
Newer version
Newer versionKimi K3.1We compared the version names in our catalogue.
🇺🇸 US
Inference by Meganova
128.5
From 401 benchmarks. Plenty of data behind it.
103 ±5
Coding IQ 102.9, 95% interval 98–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4725cache read $0.0735$2.3625Not published by the provider
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±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.
148.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.5
🇨🇳
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.
Newer version
Newer versionKimi K3.1We compared the version names in our catalogue.
moonshotai/kimi-k2.5:thinking · thinking · 🇺🇸 US
Inference by Meganova
128.5
From 401 benchmarks. Plenty of data behind it.
103 ±5
Coding IQ 102.9, 95% interval 98–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4725cache read $0.0735$2.3625Not published by the provider
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±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.
148.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.5
🇨🇳
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.
Newer version
Newer versionKimi K3.1We compared the version names in our catalogue.
🇺🇸 US
Inference by Novita AI
128.5
From 401 benchmarks. Plenty of data behind it.
103 ±5
Coding IQ 102.9, 95% interval 98–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.105$3.15Not published by the provider
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±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.
148.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.5
🇨🇳
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.
Newer version
Newer versionKimi K3.1We compared the version names in our catalogue.
moonshotai/kimi-k2.5:thinking · thinking · 🇺🇸 US
Inference by Novita AI
128.5
From 401 benchmarks. Plenty of data behind it.
103 ±5
Coding IQ 102.9, 95% interval 98–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.105$3.15Not published by the provider
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±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.
148.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.5
🇨🇳
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.
Newer version
Newer versionKimi K3.1We compared the version names in our catalogue.
int4 · 🇸🇬 SG
Inference by SiliconFlow
128.5
From 401 benchmarks. Plenty of data behind it.
103 ±5
Coding IQ 102.9, 95% interval 98–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.525cache read $0.1575$2.625Not published by the provider
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±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.
148.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.5
🇨🇳
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.
Newer version
Newer versionKimi K3.1We compared the version names in our catalogue.
moonshotai/kimi-k2.5:thinking · thinking · int4 · 🇸🇬 SG
Inference by SiliconFlow
128.5
From 401 benchmarks. Plenty of data behind it.
103 ±5
Coding IQ 102.9, 95% interval 98–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.525cache read $0.1575$2.625Not published by the provider
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±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.
148.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.7 Flash
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Newer version
Newer versionQwen3.8 FlashWe compared release dates and version names in our catalogue.
128.4
From 15 benchmarks. A fair amount of data behind it.
Not enough data yet$0.03cache read $0.006$0.13cache write $0.038992KNot enough data yetNot measuredNot measured
Qwen3.7 Flash
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Newer version
Newer versionQwen3.8 FlashWe compared release dates and version names in our catalogue.
qwen/qwen3.7-flash:thinking · thinking
128.4
From 15 benchmarks. A fair amount of data behind it.
Not enough data yet$0.03cache read $0.006$0.13cache write $0.038984KNot enough data yetNot measuredNot measured
Claude Opus 4.1
🇺🇸
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.
128.3
From 249 benchmarks. Plenty of data behind it.
109 ±7
Coding IQ 108.6, 95% interval 102–115. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$15.00cache read $1.50$75.00cache write $18.75200K
34.6
Poor value (34.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
35 ±9
Terse (35 ±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.
144.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.1
🇺🇸
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-4.1:thinking · thinking
128.3
From 249 benchmarks. Plenty of data behind it.
109 ±7
Coding IQ 108.6, 95% interval 102–115. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$15.00cache read $1.50$75.00cache write $18.75200K
34.6
Poor value (34.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
35 ±9
Terse (35 ±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.
144.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.1
🇺🇸
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-4.1:thinking:1024 · thinking+1024
128.3
From 249 benchmarks. Plenty of data behind it.
109 ±7
Coding IQ 108.6, 95% interval 102–115. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$15.00cache read $1.50$75.00cache write $18.75200K
34.6
Poor value (34.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
35 ±9
Terse (35 ±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.
144.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.1
🇺🇸
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-4.1:thinking:32768 · thinking+32768
128.3
From 249 benchmarks. Plenty of data behind it.
109 ±7
Coding IQ 108.6, 95% interval 102–115. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$15.00cache read $1.50$75.00cache write $18.75200K
34.6
Poor value (34.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
35 ±9
Terse (35 ±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.
144.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Claude Opus 4.1
🇺🇸
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-4.1:thinking:8192 · thinking+8192
128.3
From 249 benchmarks. Plenty of data behind it.
109 ±7
Coding IQ 108.6, 95% interval 102–115. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$15.00cache read $1.50$75.00cache write $18.75200K
34.6
Poor value (34.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
35 ±9
Terse (35 ±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.
144.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.11cache read $0.05$0.33262K
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
tee
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.15cache read $0.075$0.46262K
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
TEE/gemma4-31b:thinking · tee+thinking
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.12cache read $0.012$0.37262K
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.10$0.34262K
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
NVFP4 · 🇰🇷 KR
Inference by Aoru
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105cache read $0.0525$0.4725Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
🇮🇩 ID
Inference by Arliai
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.189$0.189Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105$0.357Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · fp4 · 🇺🇸 US
Inference by Coreweave
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.105$0.357Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
nvfp4 · GLB
Inference by Creekside
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0954cache read $0.0112$0.2806Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · nvfp4 · GLB
Inference by Creekside
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0954cache read $0.0112$0.2806Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Crusoe
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · bf16 · 🇺🇸 US
Inference by Crusoe
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 DeepInfra
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.21cache read $0.021$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Friendli AI
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · 🇺🇸 US
Inference by Friendli AI
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1559cache read $0.0606$0.441Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 · 🇦🇪 AE
Inference by Hyperfusion
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1365$0.399Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · bf16 · 🇦🇪 AE
Inference by Hyperfusion
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1365$0.399Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
int4 · GLB
Inference by Kitani
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty 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$0.4515Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Modal
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.357cache read $0.105$0.105Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · 🇺🇸 US
Inference by Modal
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.357cache read $0.105$0.105Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Modelrun
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.7875cache read $0.21$1.05Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · fp4 · 🇺🇸 US
Inference by Modelrun
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.7875cache read $0.21$1.05Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Neuralwatt
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1512cache read $0.0151$0.441Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · fp4 · 🇺🇸 US
Inference by Neuralwatt
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1512cache read $0.0151$0.441Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 · 🇪🇺 EU
Inference by Nextbit
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1313cache read $0.0788$0.357Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · fp4 · 🇪🇺 EU
Inference by Nextbit
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1313cache read $0.0788$0.357Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Novita AI
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · bf16 · 🇺🇸 US
Inference by Novita AI
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147cache read $0.0525$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · UNK
Inference by Ollama
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147cache read $0.0525$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1575cache read $0.063$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · fp8 · 🇺🇸 US
Inference by Parasail
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1575cache read $0.063$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 · UNK
Inference by Primeintellect
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1082cache read $0.0546$0.1817Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · fp8 · UNK
Inference by Primeintellect
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1082cache read $0.0546$0.1817Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Runware
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0964cache read $0.0113$0.2807Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · UNK
Inference by Runware
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0964cache read $0.0113$0.2807Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
asap · fp4 · GLB
Inference by Sail Research
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147cache read $0.0735$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · asap+thinking · fp4 · GLB
Inference by Sail Research
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.147cache read $0.0735$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
standard · fp4 · GLB
Inference by Sail Research
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0735cache read $0.0525$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · standard+thinking · fp4 · GLB
Inference by Sail Research
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.0735cache read $0.0525$0.42Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 SambaNova
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.399$1.2075Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · 🇺🇸 US
Inference by SambaNova
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.399$1.2075Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Scx
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315$0.9555Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · bf16 · GLB
Inference by Scx
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315$0.9555Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.7875cache read $0.2625$1.05Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · fp8 · 🇸🇬 SG
Inference by SiliconFlow
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.7875cache read $0.2625$1.05Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Tinfoil
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42$1.05Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
TEE/gemma4-31b · tee · 🇺🇸 US
Inference by Tinfoil
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42$1.05Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
TEE/gemma4-31b:thinking · tee+thinking · 🇺🇸 US
Inference by Tinfoil
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42$1.05Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · 🇺🇸 US
Inference by Tinfoil
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42$1.05Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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 Uomi
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.3056$0.5316Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 31B IT
🇺🇸
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.
google/gemma-4-31b-it:thinking · thinking · fp8 · GLB
Inference by Uomi
128.3
From 333 benchmarks. Plenty of data behind it.
102 ±7
Coding IQ 101.7, 95% interval 94–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.3056$0.5316Not published by the provider
61.9
Good value (61.9). Cheaper per task than its benchmark results predict. Plenty of data behind it.
43 ±9
Terse (43 ±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.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Muse Glimmer 30B
🇺🇸
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.
128.3
From 58 benchmarks. Plenty of data behind it.
Not enough data yet$0.35cache read $0.04$1.50131K
51.7
Fair value (51.7). Costs per task about what its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Muse Glimmer 30B
🇺🇸
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.
nvfp4 · GLB
Inference by Creekside
128.3
From 58 benchmarks. Plenty of data behind it.
Not enough data yet$0.2992cache read $0.0399$1.197Not published by the provider
51.7
Fair value (51.7). Costs per task about what its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Muse Glimmer 30B
🇺🇸
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
128.3
From 58 benchmarks. Plenty of data behind it.
Not enough data yet$0.315cache read $0.042$1.26Not published by the provider
51.7
Fair value (51.7). Costs per task about what its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Muse Glimmer 30B
🇺🇸
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 Modelrun
128.3
From 58 benchmarks. Plenty of data behind it.
Not enough data yet$0.315cache read $0.042$1.26Not published by the provider
51.7
Fair value (51.7). Costs per task about what its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Muse Glimmer 30B
🇺🇸
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 Together AI
128.3
From 58 benchmarks. Plenty of data behind it.
Not enough data yet$0.3675cache read $0.042$1.575Not published by the provider
51.7
Fair value (51.7). Costs per task about what its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
OpenAI o3
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
openai/o3
128.3
From 343 benchmarks. Plenty of data behind it.
100 ±6
Coding IQ 99.6, 95% interval 93–106. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.00cache read $1.00$8.00200K
54.5
Good value (54.5). Cheaper per task than its benchmark results predict. Plenty of data behind it.
58 ±9
Chatty (58 ±9). Generates more tokens than models with similar coding results. Measured at medium effort. A fair amount of data behind it.
147
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GPT-5.4 mini
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
128.1
From 307 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 104.3, 95% interval 99–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.75cache read $0.075$4.50400K
39.8
Poor value (39.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
46 ±18
Typical (46 ±18). 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.
149.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiMo V2.5 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.
Newer version
Newer versionMiMo V2.6 ProWe compared release dates and version names in our catalogue.
128.0
From 172 benchmarks. Plenty of data behind it.
107 ±6
Coding IQ 106.5, 95% interval 101–112. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.48cache read $0.096$1.80262K
59.5
Good value (59.5). 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.
Not measured
MiMo V2.5 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.
Newer version
Newer versionMiMo V2.6 ProWe compared release dates and version names in our catalogue.
xiaomi/mimo-v2.5-pro:thinking · thinking
128.0
From 172 benchmarks. Plenty of data behind it.
107 ±6
Coding IQ 106.5, 95% interval 101–112. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.48cache read $0.096$1.80262K
59.5
Good value (59.5). 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.
Not measured
MiMo V2.5 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.
Newer version
Newer versionMiMo V2.6 ProWe compared release dates and version names in our catalogue.
🇺🇸 US
Inference by Digitalocean
128.0
From 172 benchmarks. Plenty of data behind it.
107 ±6
Coding IQ 106.5, 95% interval 101–112. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.54cache read $0.11$2.00Not published by the provider
59.5
Good value (59.5). 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.
Not measured
MiMo V2.5 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.
Newer version
Newer versionMiMo V2.6 ProWe compared release dates and version names in our catalogue.
xiaomi/mimo-v2.5-pro:thinking · thinking · 🇺🇸 US
Inference by Digitalocean
128.0
From 172 benchmarks. Plenty of data behind it.
107 ±6
Coding IQ 106.5, 95% interval 101–112. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.54cache read $0.11$2.00Not published by the provider
59.5
Good value (59.5). 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.
Not measured
MiMo V2.5 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.
Newer version
Newer versionMiMo V2.6 ProWe compared release dates and version names in our catalogue.
🇺🇸 US
Inference by Gmicloud
128.0
From 172 benchmarks. Plenty of data behind it.
107 ±6
Coding IQ 106.5, 95% interval 101–112. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4568cache read $0.0042$0.9135Not published by the provider
59.5
Good value (59.5). 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.
Not measured
MiMo V2.5 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.
Newer version
Newer versionMiMo V2.6 ProWe compared release dates and version names in our catalogue.
xiaomi/mimo-v2.5-pro:thinking · thinking · 🇺🇸 US
Inference by Gmicloud
128.0
From 172 benchmarks. Plenty of data behind it.
107 ±6
Coding IQ 106.5, 95% interval 101–112. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4568cache read $0.0042$0.9135Not published by the provider
59.5
Good value (59.5). 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.
Not measured
Gemini 3.5 Flash Lite
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
127.5
From 169 benchmarks. Plenty of data behind it.
105 ±6
Coding IQ 105.2, 95% interval 99–111. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.03$2.501M
46.2
Poor value (46.2). Costs more per task than its benchmark results predict. Plenty of data behind it.
50 ±9
Typical (50 ±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.
145.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling Small
🇺🇸
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.
127.4
From 169 benchmarks. Plenty of data behind it.
107 ±7
Coding IQ 107.0, 95% interval 101–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.45cache read $0.10$1.20524K
45.5
Poor value (45.5). Costs more per task than its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±9). Generates fewer tokens than models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
150.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling Small
🇺🇸
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.
thinkingmachines/Inkling-Small:thinking · thinking
127.4
From 169 benchmarks. Plenty of data behind it.
107 ±7
Coding IQ 107.0, 95% interval 101–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.45cache read $0.10$1.20524K
45.5
Poor value (45.5). Costs more per task than its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±9). Generates fewer tokens than models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
150.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling Small
🇺🇸
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 DeepInfra
127.4
From 169 benchmarks. Plenty of data behind it.
107 ±7
Coding IQ 107.0, 95% interval 101–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4725cache read $0.105$1.26Not published by the provider
45.5
Poor value (45.5). Costs more per task than its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±9). Generates fewer tokens than models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
150.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling Small
🇺🇸
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.
thinkingmachines/Inkling-Small:thinking · thinking · fp4 · 🇺🇸 US
Inference by DeepInfra
127.4
From 169 benchmarks. Plenty of data behind it.
107 ±7
Coding IQ 107.0, 95% interval 101–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4725cache read $0.105$1.26Not published by the provider
45.5
Poor value (45.5). Costs more per task than its benchmark results predict. Plenty of data behind it.
42 ±9
Terse (42 ±9). Generates fewer tokens than models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
150.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Hy3 (Deep Infra)
🇨🇳
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.
tencent/hy3
127.2
From 119 benchmarks. Plenty of data behind it.
111 ±6
Coding IQ 111.3, 95% interval 105–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.066cache read $0.029$0.26262K
52.3
Fair value (52.3). Costs per task about what its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4004cache read $0.0801$1.9292200K
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original · original
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4004cache read $0.0801$1.9292200K
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original:thinking · original+thinking
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4004cache read $0.0801$1.9292200K
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7:thinking · thinking
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4004cache read $0.0801$1.9292200K
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
fp4 · 🇺🇸 US
Inference by DeepInfra
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42cache read $0.084$1.8375Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original · original · fp4 · 🇺🇸 US
Inference by DeepInfra
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42cache read $0.084$1.8375Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original:thinking · original+thinking · fp4 · 🇺🇸 US
Inference by DeepInfra
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42cache read $0.084$1.8375Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7:thinking · thinking · fp4 · 🇺🇸 US
Inference by DeepInfra
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42cache read $0.084$1.8375Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
fp4
Inference by Mancer
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.735cache read $0.3675$2.625Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original · original · fp4
Inference by Mancer
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.735cache read $0.3675$2.625Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original:thinking · original+thinking · fp4
Inference by Mancer
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.735cache read $0.3675$2.625Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7:thinking · thinking · fp4
Inference by Mancer
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.735cache read $0.3675$2.625Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
fp8 · 🇺🇸 US
Inference by Meganova
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. 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
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original · original · fp8 · 🇺🇸 US
Inference by Meganova
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. 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
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original:thinking · original+thinking · fp8 · 🇺🇸 US
Inference by Meganova
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. 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
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7:thinking · thinking · fp8 · 🇺🇸 US
Inference by Meganova
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. 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
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
fp8 · 🇺🇸 US
Inference by Novita AI
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1155$2.31Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original · original · fp8 · 🇺🇸 US
Inference by Novita AI
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1155$2.31Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original:thinking · original+thinking · fp8 · 🇺🇸 US
Inference by Novita AI
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1155$2.31Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7:thinking · thinking · fp8 · 🇺🇸 US
Inference by Novita AI
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1155$2.31Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
bf16 · GLB
Inference by Uomi
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.6615$2.4255Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original · original · bf16 · GLB
Inference by Uomi
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.6615$2.4255Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original:thinking · original+thinking · bf16 · GLB
Inference by Uomi
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.6615$2.4255Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7:thinking · thinking · bf16 · GLB
Inference by Uomi
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.6615$2.4255Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
fp4 · 🇺🇸 US
Inference by Venice
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4204cache read $0.0841$2.0257Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original · original · fp4 · 🇺🇸 US
Inference by Venice
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4204cache read $0.0841$2.0257Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original:thinking · original+thinking · fp4 · 🇺🇸 US
Inference by Venice
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4204cache read $0.0841$2.0257Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7:thinking · thinking · fp4 · 🇺🇸 US
Inference by Venice
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.4204cache read $0.0841$2.0257Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
fp4 · 🇸🇬 SG
Inference by Z.ai
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1155$2.31Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original · original · fp4 · 🇸🇬 SG
Inference by Z.ai
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1155$2.31Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7-original:thinking · original+thinking · fp4 · 🇸🇬 SG
Inference by Z.ai
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1155$2.31Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.7
🇨🇳
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.
Newer version
Newer versionGLM 5.3We compared release dates and version names in our catalogue.
z-ai/glm-4.7:thinking · thinking · fp4 · 🇸🇬 SG
Inference by Z.ai
127.0
From 234 benchmarks. Plenty of data behind it.
104 ±5
Coding IQ 103.5, 95% interval 98–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1155$2.31Not published by the provider
41.0
Poor value (41.0). 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.
143.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiMo V2.5
🇨🇳
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.
127.0
From 145 benchmarks. Plenty of data behind it.
108 ±6
Coding IQ 107.8, 95% interval 102–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.422cache read $0.0844$2.111M
59.9
Good value (59.9). 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.
Not measured
MiMo V2.5
🇨🇳
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.
xiaomi/mimo-v2.5:thinking · thinking
127.0
From 145 benchmarks. Plenty of data behind it.
108 ±6
Coding IQ 107.8, 95% interval 102–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.422cache read $0.0844$2.111M
59.9
Good value (59.9). 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.
Not measured
MiMo V2.5
🇨🇳
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 Uomi
127.0
From 145 benchmarks. Plenty of data behind it.
108 ±6
Coding IQ 107.8, 95% interval 102–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.422$2.11Not published by the provider
59.9
Good value (59.9). 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.
Not measured
MiMo V2.5
🇨🇳
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.
xiaomi/mimo-v2.5:thinking · thinking · bf16 · GLB
Inference by Uomi
127.0
From 145 benchmarks. Plenty of data behind it.
108 ±6
Coding IQ 107.8, 95% interval 102–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.422$2.11Not published by the provider
59.9
Good value (59.9). 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.
Not measured
MiMo V2.5
🇨🇳
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 · 🇺🇸 US
Inference by Venice
127.0
From 145 benchmarks. Plenty of data behind it.
108 ±6
Coding IQ 107.8, 95% interval 102–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.422cache read $0.0844$2.11Not published by the provider
59.9
Good value (59.9). 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.
Not measured
MiMo V2.5
🇨🇳
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.
xiaomi/mimo-v2.5:thinking · thinking · fp8 · 🇺🇸 US
Inference by Venice
127.0
From 145 benchmarks. Plenty of data behind it.
108 ±6
Coding IQ 107.8, 95% interval 102–114. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.422cache read $0.0844$2.11Not published by the provider
59.9
Good value (59.9). 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.
Not measured
Inkling FP4
🇺🇸
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.
thinkingmachines/inkling
126.9
From 215 benchmarks. Plenty of data behind it.
102 ±6
Coding IQ 101.9, 95% interval 96–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.17$4.051M
27.9
Poor value (27.9). Costs more per task than its benchmark results predict. Plenty of data behind it.
58 ±9
Chatty (58 ±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.
148.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling FP4
🇺🇸
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.
thinkingmachines/inkling:thinking · thinking
126.9
From 215 benchmarks. Plenty of data behind it.
102 ±6
Coding IQ 101.9, 95% interval 96–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.00cache read $0.17$4.051M
27.9
Poor value (27.9). Costs more per task than its benchmark results predict. Plenty of data behind it.
58 ±9
Chatty (58 ±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.
148.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling FP4
🇺🇸
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.
thinkingmachines/inkling · 🇺🇸 US
Inference by Fireworks AI
126.9
From 215 benchmarks. Plenty of data behind it.
102 ±6
Coding IQ 101.9, 95% interval 96–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.05cache read $0.1785$4.2525Not published by the provider
27.9
Poor value (27.9). Costs more per task than its benchmark results predict. Plenty of data behind it.
58 ±9
Chatty (58 ±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.
148.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling FP4
🇺🇸
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.
thinkingmachines/inkling:thinking · thinking · 🇺🇸 US
Inference by Fireworks AI
126.9
From 215 benchmarks. Plenty of data behind it.
102 ±6
Coding IQ 101.9, 95% interval 96–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.05cache read $0.1785$4.2525Not published by the provider
27.9
Poor value (27.9). Costs more per task than its benchmark results predict. Plenty of data behind it.
58 ±9
Chatty (58 ±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.
148.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling FP4
🇺🇸
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.
thinkingmachines/inkling · 🇺🇸 US
Inference by Together AI
126.9
From 215 benchmarks. Plenty of data behind it.
102 ±6
Coding IQ 101.9, 95% interval 96–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.05cache read $0.1785$4.2525Not published by the provider
27.9
Poor value (27.9). Costs more per task than its benchmark results predict. Plenty of data behind it.
58 ±9
Chatty (58 ±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.
148.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Inkling FP4
🇺🇸
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.
thinkingmachines/inkling:thinking · thinking · 🇺🇸 US
Inference by Together AI
126.9
From 215 benchmarks. Plenty of data behind it.
102 ±6
Coding IQ 101.9, 95% interval 96–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.05cache read $0.1785$4.2525Not published by the provider
27.9
Poor value (27.9). Costs more per task than its benchmark results predict. Plenty of data behind it.
58 ±9
Chatty (58 ±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.
148.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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.
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60cache read $0.12$3.00262K
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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-k2.7-code-highspeed · highspeed
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.90cache read $0.32$8.00262K
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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 Alibaba Cloud
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.9975cache read $0.1995$4.20Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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.
int4 · 🇺🇸 US
Inference by Coreweave
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.126$3.15Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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.
int4 · 🇪🇺 EU
Inference by Inceptron
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.6405cache read $0.1785$3.7275Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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 Meganova
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.7875cache read $0.1575$3.675Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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.
int4 · 🇺🇸 US
Inference by Novita AI
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.9975cache read $0.1995$4.20Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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.
UNK
Inference by Ollama
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.9975cache read $0.1995$4.20Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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 · GLB
Inference by Scx
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.7739cache read $0.1502$3.234Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.9021cache read $0.1889$3.99Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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.
🇨🇳 CN
Inference by Streamlake
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.7481cache read $0.1496$3.15Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2.7 Code
🇨🇳
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.
int4 · 🇺🇸 US
Inference by Venice
126.9
From 102 benchmarks. Plenty of data behind it.
105 ±9
Coding IQ 104.5, 95% interval 96–113. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.7875cache read $0.168$3.675Not published by the provider
43.6
Poor value (43.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
62 ±9
Chatty (62 ±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.
151.2
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 5 Turbo
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
126.8
From 54 benchmarks. A fair amount of data behind it.
≈ 104 ±11
Coding IQ 103.5, 95% interval 93–114. 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.
$1.20cache read $0.24$4.00203K
48.2
Fair value (48.2). Costs per task about what its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
GLM 5V Turbo
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
126.8
From 68 benchmarks. Plenty of data behind it.
102 ±22
Coding IQ 101.9, 95% interval 80–124. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.20cache read $0.24$4.00203K
52.2
Fair value (52.2). Costs per task about what its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
GLM 5V Turbo
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
z-ai/glm-5v-turbo:thinking · thinking
126.8
From 68 benchmarks. Plenty of data behind it.
102 ±22
Coding IQ 101.9, 95% interval 80–124. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$1.20cache read $0.24$4.00203K
52.2
Fair value (52.2). Costs per task about what its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.06$1.20512K
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30cache read $0.06$1.20512K
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
fp4 · 🇺🇸 US
Inference by Coreweave
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2415cache read $0.0525$1.008Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · fp4 · 🇺🇸 US
Inference by Coreweave
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2415cache read $0.0525$1.008Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
fp8 · 🇺🇸 US
Inference by DeepInfra
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.294cache read $0.0588$1.155Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · fp8 · 🇺🇸 US
Inference by DeepInfra
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.294cache read $0.0588$1.155Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · fp8 · UNK
Inference by Featherless AI
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.5775cache read $0.063$2.31Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · fp8 · 🇺🇸 US
Inference by Gmicloud
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.252cache read $0.0504$1.008Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
🇺🇸 US
Inference by Novita AI
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315cache read $0.063$1.26Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · 🇺🇸 US
Inference by Novita AI
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315cache read $0.063$1.26Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
UNK
Inference by Ollama
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.126$2.52Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · UNK
Inference by Ollama
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.126$2.52Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
fp8 · 🇺🇸 US
Inference by Parasail
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315cache read $0.063$1.26Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · fp8 · 🇺🇸 US
Inference by Parasail
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315cache read $0.063$1.26Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · bf16 · 🇺🇸 US
Inference by Runanywhere
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315cache read $0.063$1.26Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · UNK
Inference by Runware
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2835cache read $0.0567$1.134Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
fp8 · 🇸🇬 SG
Inference by SiliconFlow
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.126$2.52Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · fp8 · 🇸🇬 SG
Inference by SiliconFlow
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.126$2.52Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
fp8 · 🇺🇸 US
Inference by Venice
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315cache read $0.063$1.26Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
MiniMax M3
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
minimax/minimax-m3:thinking · thinking · fp8 · 🇺🇸 US
Inference by Venice
126.8
From 332 benchmarks. Plenty of data behind it.
107 ±9
Coding IQ 107.2, 95% interval 98–117. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315cache read $0.063$1.26Not published by the provider
50.9
Fair value (50.9). Costs per task about what its benchmark results predict. Plenty of data behind it.
71 ±8
Chatty (71 ±8). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
147.8
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 Plus
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Newer version
Newer versionQwen3.7 PlusWe compared release dates and version names in our catalogue.
126.8
From 53 benchmarks. Plenty of data behind it.
104 ±7
Coding IQ 103.9, 95% interval 97–111. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.40cache read $0.04$2.40cache write $0.50984K
51.4
Fair value (51.4). Costs per task about what its benchmark results predict. Plenty of data behind it.
46 ±9
Typical (46 ±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.
Not measured
Qwen3.5 Plus
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Newer version
Newer versionQwen3.7 PlusWe compared release dates and version names in our catalogue.
qwen/qwen3.5-plus:thinking · thinking
126.8
From 53 benchmarks. Plenty of data behind it.
104 ±7
Coding IQ 103.9, 95% interval 97–111. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.40cache read $0.04$2.40cache write $0.50984K
51.4
Fair value (51.4). Costs per task about what its benchmark results predict. Plenty of data behind it.
46 ±9
Typical (46 ±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.
Not measured
Qwen3.5 122B-A10B
🇨🇳
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.
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.26cache read $0.13$2.08131K
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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/qwen3.5-122b-a10b:thinking · thinking
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.26cache read $0.13$2.08131K
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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.
🇸🇬 SG
Inference by Alibaba Cloud
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42$3.36Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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/qwen3.5-122b-a10b:thinking · thinking · 🇸🇬 SG
Inference by Alibaba Cloud
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42$3.36Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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 Atlascloud
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315$2.52Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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/qwen3.5-122b-a10b:thinking · thinking · 🇺🇸 US
Inference by Atlascloud
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315$2.52Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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 Novita AI
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42$3.36Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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/qwen3.5-122b-a10b:thinking · thinking · bf16 · 🇺🇸 US
Inference by Novita AI
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.42$3.36Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.273$2.184Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3.5 122B-A10B
🇨🇳
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/qwen3.5-122b-a10b:thinking · thinking · fp8 · 🇸🇬 SG
Inference by SiliconFlow
126.7
From 287 benchmarks. Plenty of data behind it.
99 ±7
Coding IQ 99.3, 95% interval 92–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.273$2.184Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Particle 1
🇺🇸
Maker headquartered in US, according to the company's own website. The citation is on the model page.
≈ 126
From 3 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet$0.25cache read $0.03$1.001MNot enough data yetNot measuredNot measured
Nemotron 3.5 Super VL
🇺🇸
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.
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.15cache read $0.02$0.50262K
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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 Baseten
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.21cache read $0.042$0.525Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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.
nvfp4 · 🇺🇸 US
Inference by Coreweave
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.1575cache read $0.021$0.4935Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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 Crusoe
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.105cache read $0.084$0.42Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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.
nvfp4 · GLB
Inference by Darkbloom
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.063cache read $0.0394$0.3675Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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 DeepInfra
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.1575cache read $0.021$0.525Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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.
nvfp4 · 🇺🇸 US
Inference by Ionet
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.0945cache read $0.0126$0.315Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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 Nebius AI Studio
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.315cache read $0.063$0.945Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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.
nvfp4 · UNK
Inference by Primeintellect
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.1575cache read $0.021$0.4725Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Nemotron 3.5 Super VL
🇺🇸
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 Together AI
125.9
From 18 benchmarks. A fair amount of data behind it.
Not enough data yet$0.1575cache read $0.021$0.525Not published by the provider
60.2
Good value (60.2). Cheaper per task than its benchmark results predict. A fair amount of data behind it.
54 ±9
Typical (54 ±9). Generates about as many tokens as models with similar coding results. Measured with reasoning off. Little data behind it, so treat it as a rough estimate.
Not measured
Ling 3.0 Flash
🇨🇳
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.
Newer version
Newer versionLing 3.2 FlashWe compared the version names in our catalogue.
125.6
From 52 benchmarks. Plenty of data behind it.
Not enough data yet$0.075cache read $0.015$0.22262K
64.2
Good value (64.2). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Ling 3.0 Flash
🇨🇳
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.
Newer version
Newer versionLing 3.2 FlashWe compared the version names in our catalogue.
inclusionai/ling-3.0-flash:thinking · thinking
125.6
From 52 benchmarks. Plenty of data behind it.
Not enough data yet$0.075cache read $0.015$0.22262K
64.2
Good value (64.2). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.20cache read $0.02$1.75260K
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
qwen/qwen3.5-27b:thinking · thinking
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.20cache read $0.02$1.75260K
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
🇸🇬 SG
Inference by Alibaba Cloud
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315$2.52Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
qwen/qwen3.5-27b:thinking · thinking · 🇸🇬 SG
Inference by Alibaba Cloud
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315$2.52Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
🇺🇸 US
Inference by Atlascloud
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2835$2.268Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
qwen/qwen3.5-27b:thinking · thinking · 🇺🇸 US
Inference by Atlascloud
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2835$2.268Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
fp8 · 🇺🇸 US
Inference by Chutes
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.21cache read $0.021$1.8375Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
qwen/qwen3.5-27b:thinking · thinking · fp8 · 🇺🇸 US
Inference by Chutes
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.21cache read $0.021$1.8375Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
fp8 · 🇺🇸 US
Inference by DeepInfra
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.273$2.73Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
qwen/qwen3.5-27b:thinking · thinking · fp8 · 🇺🇸 US
Inference by DeepInfra
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.273$2.73Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
bf16 · 🇺🇸 US
Inference by Novita AI
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315$2.52Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
qwen/qwen3.5-27b:thinking · thinking · bf16 · 🇺🇸 US
Inference by Novita AI
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.315$2.52Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
fp8 · 🇸🇬 SG
Inference by SiliconFlow
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2625$2.10Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured
Qwen3.5 27B
🇨🇳
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.
Newer version
Newer versionQwen 3.8 27BWe compared release dates and version names in our catalogue.
qwen/qwen3.5-27b:thinking · thinking · fp8 · 🇸🇬 SG
Inference by SiliconFlow
125.6
From 331 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 100.7, 95% interval 94–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.2625$2.10Not published by the provider
42.4
Poor value (42.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
56 ±9
Chatty (56 ±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.
Not measured

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

How this data works