Skip to content

Listings

All listings: Novita AI

Every listing for the models on page 3 of 8 of Novita AI, including each model's other versions and routes. Back to Novita AI

Sold by Novita AI

29 listings on this page sold by Novita AI.

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
Qwen3 Max
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Newer version
Newer versionQwen4 MaxWe compared release dates and version names in our catalogue.
125.2
From 157 benchmarks. Plenty of data behind it.
96 ±10
Coding IQ 96.3, 95% interval 86–107. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$2.11$8.45262K
43.3
Poor value (43.3). Costs more per task than its benchmark results predict. Plenty of data behind it.
41 ±9
Terse (41 ±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.
142.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3.2
🇨🇳
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 versionDeepSeek V4We compared the version names in our catalogue.
fp8
125.0
From 247 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 101.2, 95% interval 94–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.269cache read $0.1345$0.40164K
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
146.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 26B A4B
🇺🇸
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-4-26b-a4b-it · bf16
124.9
From 216 benchmarks. Plenty of data behind it.
102 ±8
Coding IQ 101.7, 95% interval 94–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.13$0.40262K
56.4
Good value (56.4). Cheaper per task than its benchmark results predict. A fair amount 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.
141.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Ling 3.0 Flash VL
🇨🇳
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.
≈ 125
From 3 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet$0.075cache read $0.015$0.22262K
65.3
Good value (65.3). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Qwen 3 235B A22B Thinking 2507
🇨🇳
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.
124.3
From 161 benchmarks. Plenty of data behind it.
100 ±8stale
Coding IQ 99.6, 95% interval 91–108. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. Its newest result is over a year old.
$0.30$3.00131K
40.6
Poor value (40.6). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Kimi K2 Thinking
🇨🇳
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
123.8
From 184 benchmarks. Plenty of data behind it.
100 ±6
Coding IQ 100.2, 95% interval 94–106. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60cache read $0.15$2.50262K
41.6
Poor value (41.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
50 ±18
Typical (50 ±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.
145.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3.2 Exp
🇨🇳
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.
123.7
From 179 benchmarks. Plenty of data behind it.
98 ±7
Coding IQ 97.5, 95% interval 90–105. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.27$0.41164K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
144.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.6 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.
123.4
From 160 benchmarks. Plenty of data behind it.
101 ±6
Coding IQ 100.9, 95% interval 95–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60$3.60262K
39.6
Poor value (39.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
146.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.6
🇨🇳
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
123.1
From 174 benchmarks. Plenty of data behind it.
96 ±22
Coding IQ 96.0, 95% interval 74–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.55cache read $0.11$2.20205K
46.8
Poor value (46.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
MiniMax M2.5
🇨🇳
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.
122.8
From 192 benchmarks. Plenty of data behind it.
100 ±6
Coding IQ 100.4, 95% interval 94–106. 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$1.20205K
48.7
Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it.
64 ±9
Chatty (64 ±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
MiniMax M2.5
🇨🇳
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.
novita/minimax/minimax-m2.5-highspeed · highspeed
122.8
From 192 benchmarks. Plenty of data behind it.
100 ±6
Coding IQ 100.4, 95% interval 94–106. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60cache read $0.03$2.40205K
48.7
Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it.
64 ±9
Chatty (64 ±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
MiniMax M2.1
🇨🇳
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.
122.3
From 97 benchmarks. Plenty of data behind it.
97 ±6
Coding IQ 96.5, 95% interval 91–102. 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$1.20205K
45.1
Poor value (45.1). Costs more per task than its benchmark results predict. Plenty of data behind it.
59 ±13
Chatty (59 ±13). 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
DeepSeek R1
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
121.9
From 412 benchmarks. Plenty of data behind it.
93 ±14
Coding IQ 92.9, 95% interval 79–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$4.00$4.0064K
36.1
Poor value (36.1). Costs more per task than its benchmark results predict. Plenty of data behind it.
60 ±11
Chatty (60 ±11). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
140.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek R1
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
deepseek/deepseek-r1-0528 · 0528 · fp8
121.9
From 412 benchmarks. Plenty of data behind it.
93 ±14
Coding IQ 92.9, 95% interval 79–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.70cache read $0.35$2.50164K
36.1
Poor value (36.1). Costs more per task than its benchmark results predict. Plenty of data behind it.
60 ±11
Chatty (60 ±11). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
140.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.5
🇨🇳
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 versionGLM 5.3We compared release dates and version names in our catalogue.
121.9
From 110 benchmarks. Plenty of data behind it.
95 ±7
Coding IQ 95.3, 95% interval 88–103. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60cache read $0.11$2.20131K
41.4
Poor value (41.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
DeepSeek V3.1 Terminus
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
fp8
121.4
From 57 benchmarks. Plenty of data behind it.
≈ 99 ±13
Coding IQ 99.4, 95% interval 86–112. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.27cache read $0.135$1.00131K
57.6
Good value (57.6). Cheaper per task than its benchmark results predict. A fair amount 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.
Not measured
MiniMax M1 80K
🇨🇳
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.
121.3
From 27 benchmarks. Plenty of data behind it.
Not enough data yet$0.55$2.201MNot enough data yetNot measuredNot measured
Qwen3.5 35B A3B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Newer version
Newer versionQwen3.6 35B A3BWe compared release dates and version names in our catalogue.
121.1
From 311 benchmarks. Plenty of data behind it.
95 ±8
Coding IQ 95.0, 95% interval 87–103. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.25$2.00262K
40.3
Poor value (40.3). Costs more per task than its benchmark results predict. Plenty of data behind it.
59 ±9
Chatty (59 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
ERNIE 4.5 VL 424B A47B
🇨🇳
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.
≈ 121
From 3 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet$0.42$1.25123K
54.2
Good value (54.2). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Kimi K2 Instruct
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
121.0
From 184 benchmarks. Plenty of data behind it.
93 ±6
Coding IQ 93.0, 95% interval 87–99. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.57$2.30131K
50.2
Fair value (50.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
Not measuredNot measured
Qwen3 235B A22B Instruct 2507
🇨🇳
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.
120.7
From 258 benchmarks. Plenty of data behind it.
92 ±11
Coding IQ 92.4, 95% interval 82–103. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.09$0.58131K
58.8
Good value (58.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
138
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3 VL 235B
🇨🇳
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.
novita/qwen/qwen3-vl-235b-a22b-instruct
120.4
From 170 benchmarks. Plenty of data behind it.
≈ 94 ±18
Coding IQ 94.2, 95% interval 76–113. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30$1.50131K
45.0
Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3 VL 235B
🇨🇳
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.
novita/qwen/qwen3-vl-235b-a22b-thinking · thinking
120.4
From 170 benchmarks. Plenty of data behind it.
≈ 94 ±18
Coding IQ 94.2, 95% interval 76–113. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.98$3.95131K
45.0
Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3 Next 80B A3B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
novita/qwen/qwen3-next-80b-a3b-thinking · thinking
120.2
From 191 benchmarks. Plenty of data behind it.
90 ±12
Coding IQ 90.0, 95% interval 78–102. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.15$1.50131K
51.1
Fair value (51.1). Costs per task about what its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Step 3.7 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.
120.2
From 41 benchmarks. Plenty of data behind it.
106 ±15
Coding IQ 105.6, 95% interval 90–121. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.20cache read $0.04$1.15262K
47.0
Poor value (47.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
57 ±9
Chatty (57 ±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
MiniMax M2.7
🇨🇳
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.
119.8
From 207 benchmarks. Plenty of data behind it.
98 ±12
Coding IQ 98.4, 95% interval 87–110. 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$1.20205K
50.7
Fair value (50.7). Costs per task about what 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.
Not measured
MiniMax M2.7
🇨🇳
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.
novita/minimax/minimax-m2.7-highspeed · highspeed
119.8
From 207 benchmarks. Plenty of data behind it.
98 ±12
Coding IQ 98.4, 95% interval 87–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60cache read $0.06$2.40205K
50.7
Fair value (50.7). Costs per task about what 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.
Not measured
Qwen3.6 35B A3B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
119.8
From 160 benchmarks. Plenty of data behind it.
95 ±14
Coding IQ 94.8, 95% interval 81–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.248$1.485262K
38.5
Poor value (38.5). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3.1
🇨🇳
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
119.6
From 122 benchmarks. Plenty of data behind it.
94 ±8
Coding IQ 94.3, 95% interval 86–103. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.27cache read $0.135$1.00131K
44.0
Poor value (44.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
140.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API

Run by Novita AI

49 listings on this page where Novita AI runs the model. Another provider sells these listings.

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
ModelProviderllminfo IQCoding IQInput $/1MOutput $/1MMax Input TokensValueYapEpoch Capabilities Index
DeepSeek V3.2
🇨🇳
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 versionDeepSeek V4We compared the version names in our catalogue.
 Hugging Face
125.0
From 247 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 101.2, 95% interval 94–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.269$0.40164K
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
146.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3.2
🇨🇳
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 versionDeepSeek V4We compared the version names in our catalogue.
 Llmgateway
125.0
From 247 benchmarks. Plenty of data behind it.
101 ±7
Coding IQ 101.2, 95% interval 94–108. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.269cache read $0.1345$0.40Not published by the provider
63.8
Good value (63.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
146.3
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 26B A4B
🇺🇸
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-4-26B-A4B-it
 Hugging Face
124.9
From 216 benchmarks. Plenty of data behind it.
102 ±8
Coding IQ 101.7, 95% interval 94–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.13$0.40262K
56.4
Good value (56.4). Cheaper per task than its benchmark results predict. A fair amount 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.
141.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 26B A4B
🇺🇸
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-4-26b-a4b-it
 Llmgateway
124.9
From 216 benchmarks. Plenty of data behind it.
102 ±8
Coding IQ 101.7, 95% interval 94–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.13$0.40Not published by the provider
56.4
Good value (56.4). Cheaper per task than its benchmark results predict. A fair amount 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.
141.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 26B A4B
🇺🇸
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-4-26b-a4b-it · bf16 · 🇺🇸 US
 NanoGPT
124.9
From 216 benchmarks. Plenty of data behind it.
102 ±8
Coding IQ 101.7, 95% interval 94–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1365$0.42Not published by the provider
56.4
Good value (56.4). Cheaper per task than its benchmark results predict. A fair amount 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.
141.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 26B A4B
🇺🇸
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-4-26b-a4b-it:thinking · thinking · bf16 · 🇺🇸 US
 NanoGPT
124.9
From 216 benchmarks. Plenty of data behind it.
102 ±8
Coding IQ 101.7, 95% interval 94–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.1365$0.42Not published by the provider
56.4
Good value (56.4). Cheaper per task than its benchmark results predict. A fair amount 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.
141.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Gemma 4 26B A4B
🇺🇸
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.
novita/google/gemma-4-26b-a4b-it
 Requesty
124.9
From 216 benchmarks. Plenty of data behind it.
102 ±8
Coding IQ 101.7, 95% interval 94–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.13$0.40262K
56.4
Good value (56.4). Cheaper per task than its benchmark results predict. A fair amount 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.
141.6
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Ling 3.0 Flash VL
🇨🇳
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.
 Hugging Face
≈ 125
From 3 benchmarks. Little data behind it, so treat it as a rough estimate.
Not enough data yet$0.075$0.22262K
65.3
Good value (65.3). Cheaper per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Qwen 3 235B A22B Thinking 2507
🇨🇳
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.
 Hugging Face
124.3
From 161 benchmarks. Plenty of data behind it.
100 ±8stale
Coding IQ 99.6, 95% interval 91–108. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. Its newest result is over a year old.
$0.30$3.00131K
40.6
Poor value (40.6). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Kimi K2 Thinking
🇨🇳
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.
 Hugging Face
123.8
From 184 benchmarks. Plenty of data behind it.
100 ±6
Coding IQ 100.2, 95% interval 94–106. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60$2.50262K
41.6
Poor value (41.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
50 ±18
Typical (50 ±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.
145.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2 Thinking
🇨🇳
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
 NanoGPT
123.8
From 184 benchmarks. Plenty of data behind it.
100 ±6
Coding IQ 100.2, 95% interval 94–106. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63cache read $0.1575$2.625Not published by the provider
41.6
Poor value (41.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
50 ±18
Typical (50 ±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.
145.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2 Thinking
🇨🇳
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
 OpenRouter
123.8
From 184 benchmarks. Plenty of data behind it.
100 ±6
Coding IQ 100.2, 95% interval 94–106. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60cache read $0.15$2.50262K
41.6
Poor value (41.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
50 ±18
Typical (50 ±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.
145.5
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3.2 Exp
🇨🇳
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.
 Hugging Face
123.7
From 179 benchmarks. Plenty of data behind it.
98 ±7
Coding IQ 97.5, 95% interval 90–105. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.27$0.41164K
68.3
Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
144.4
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.6 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
 NanoGPT
123.4
From 160 benchmarks. Plenty of data behind it.
101 ±6
Coding IQ 100.9, 95% interval 95–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63$3.78Not published by the provider
39.6
Poor value (39.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
146.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.6 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/qwen3.6-27b:thinking · thinking · 🇺🇸 US
 NanoGPT
123.4
From 160 benchmarks. Plenty of data behind it.
101 ±6
Coding IQ 100.9, 95% interval 95–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.63$3.78Not published by the provider
39.6
Poor value (39.6). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
146.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.6
🇨🇳
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.
 Hugging Face
123.1
From 174 benchmarks. Plenty of data behind it.
96 ±22
Coding IQ 96.0, 95% interval 74–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.55$2.20205K
46.8
Poor value (46.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
GLM 4.6
🇨🇳
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.
 Llmgateway
123.1
From 174 benchmarks. Plenty of data behind it.
96 ±22
Coding IQ 96.0, 95% interval 74–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.55cache read $0.11$2.20Not published by the provider
46.8
Poor value (46.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
GLM 4.6
🇨🇳
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 · 🇺🇸 US
 NanoGPT
123.1
From 174 benchmarks. Plenty of data behind it.
96 ±22
Coding IQ 96.0, 95% interval 74–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.5198cache read $0.104$2.079Not published by the provider
46.8
Poor value (46.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
GLM 4.6
🇨🇳
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-4.6-original · original · bf16 · 🇺🇸 US
 NanoGPT
123.1
From 174 benchmarks. Plenty of data behind it.
96 ±22
Coding IQ 96.0, 95% interval 74–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.5198cache read $0.104$2.079Not published by the provider
46.8
Poor value (46.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
GLM 4.6
🇨🇳
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-4.6:thinking · thinking · bf16 · 🇺🇸 US
 NanoGPT
123.1
From 174 benchmarks. Plenty of data behind it.
96 ±22
Coding IQ 96.0, 95% interval 74–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.5198cache read $0.104$2.079Not published by the provider
46.8
Poor value (46.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
GLM 4.6
🇨🇳
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
 OpenRouter
123.1
From 174 benchmarks. Plenty of data behind it.
96 ±22
Coding IQ 96.0, 95% interval 74–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.495cache read $0.099$1.98205K
46.8
Poor value (46.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
GLM 4.6
🇨🇳
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.
 Requesty
123.1
From 174 benchmarks. Plenty of data behind it.
96 ±22
Coding IQ 96.0, 95% interval 74–118. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60$2.20205K
46.8
Poor value (46.8). Costs more per task than its benchmark results predict. Plenty of data behind it.
48 ±9
Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
MiniMax M2.5
🇨🇳
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.
 Llmgateway
122.8
From 192 benchmarks. Plenty of data behind it.
100 ±6
Coding IQ 100.4, 95% interval 94–106. 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$1.20Not published by the provider
48.7
Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it.
64 ±9
Chatty (64 ±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
MiniMax M2.1
🇨🇳
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.
 Llmgateway
122.3
From 97 benchmarks. Plenty of data behind it.
97 ±6
Coding IQ 96.5, 95% interval 91–102. 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$1.20Not published by the provider
45.1
Poor value (45.1). Costs more per task than its benchmark results predict. Plenty of data behind it.
59 ±13
Chatty (59 ±13). 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
DeepSeek R1
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
 Hugging Face
121.9
From 412 benchmarks. Plenty of data behind it.
93 ±14
Coding IQ 92.9, 95% interval 79–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.70$2.5064K
36.1
Poor value (36.1). Costs more per task than its benchmark results predict. Plenty of data behind it.
60 ±11
Chatty (60 ±11). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
140.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek R1
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
deepseek-ai/DeepSeek-R1-0528 · 0528
 Hugging Face
121.9
From 412 benchmarks. Plenty of data behind it.
93 ±14
Coding IQ 92.9, 95% interval 79–107. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.70$2.50164K
36.1
Poor value (36.1). Costs more per task than its benchmark results predict. Plenty of data behind it.
60 ±11
Chatty (60 ±11). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it.
140.7
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
GLM 4.5
🇨🇳
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 versionGLM 5.3We compared release dates and version names in our catalogue.
 Hugging Face
121.9
From 110 benchmarks. Plenty of data behind it.
95 ±7
Coding IQ 95.3, 95% interval 88–103. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60$2.20131K
41.4
Poor value (41.4). Costs more per task than its benchmark results predict. Plenty of data behind it.
53 ±9
Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
Not measured
MiniMax M1 80K
🇨🇳
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.
 Hugging Face
121.3
From 27 benchmarks. Plenty of data behind it.
Not enough data yet$0.55$2.201MNot enough data yetNot measuredNot measured
Qwen3.5 35B A3B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Newer version
Newer versionQwen3.6 35B A3BWe compared release dates and version names in our catalogue.
 Hugging Face
121.1
From 311 benchmarks. Plenty of data behind it.
95 ±8
Coding IQ 95.0, 95% interval 87–103. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.25$2.00262K
40.3
Poor value (40.3). Costs more per task than its benchmark results predict. Plenty of data behind it.
59 ±9
Chatty (59 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 35B A3B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Newer version
Newer versionQwen3.6 35B A3BWe compared release dates and version names in our catalogue.
🇺🇸 US
 NanoGPT
121.1
From 311 benchmarks. Plenty of data behind it.
95 ±8
Coding IQ 95.0, 95% interval 87–103. 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
40.3
Poor value (40.3). Costs more per task than its benchmark results predict. Plenty of data behind it.
59 ±9
Chatty (59 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3.5 35B A3B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Newer version
Newer versionQwen3.6 35B A3BWe compared release dates and version names in our catalogue.
qwen/qwen3.5-35b-a3b:thinking · thinking · 🇺🇸 US
 NanoGPT
121.1
From 311 benchmarks. Plenty of data behind it.
95 ±8
Coding IQ 95.0, 95% interval 87–103. 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
40.3
Poor value (40.3). Costs more per task than its benchmark results predict. Plenty of data behind it.
59 ±9
Chatty (59 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Kimi K2 Instruct
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
 Hugging Face
121.0
From 184 benchmarks. Plenty of data behind it.
93 ±6
Coding IQ 93.0, 95% interval 87–99. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.57$2.30131K
50.2
Fair value (50.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
Not measuredNot measured
Kimi K2 Instruct
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
moonshotai/Kimi-K2-Instruct-0905 · 0905
 Hugging Face
121.0
From 184 benchmarks. Plenty of data behind it.
93 ±6
Coding IQ 93.0, 95% interval 87–99. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60$2.50262K
50.2
Fair value (50.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
Not measuredNot measured
Kimi K2 Instruct
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
 Requesty
121.0
From 184 benchmarks. Plenty of data behind it.
93 ±6
Coding IQ 93.0, 95% interval 87–99. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.57$2.30131K
50.2
Fair value (50.2). Costs per task about what its benchmark results predict. Plenty of data behind it.
Not measuredNot measured
Qwen3 235B A22B Instruct 2507
🇨🇳
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.
 Hugging Face
120.7
From 258 benchmarks. Plenty of data behind it.
92 ±11
Coding IQ 92.4, 95% interval 82–103. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.09$0.58131K
58.8
Good value (58.8). Cheaper per task than its benchmark results predict. Plenty of data behind it.
Not measured
138
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
Qwen3 VL 235B
🇨🇳
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-VL-235B-A22B-Instruct
 Hugging Face
120.4
From 170 benchmarks. Plenty of data behind it.
≈ 94 ±18
Coding IQ 94.2, 95% interval 76–113. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30$1.50131K
45.0
Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3 VL 235B
🇨🇳
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-VL-235B-A22B-Thinking · thinking
 Hugging Face
120.4
From 170 benchmarks. Plenty of data behind it.
≈ 94 ±18
Coding IQ 94.2, 95% interval 76–113. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.98$3.95131K
45.0
Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3 VL 235B
🇨🇳
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-vl-235b-a22b-instruct
 Llmgateway
120.4
From 170 benchmarks. Plenty of data behind it.
≈ 94 ±18
Coding IQ 94.2, 95% interval 76–113. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.30$1.50Not published by the provider
45.0
Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it.
Not measuredNot measured
Qwen3 Next 80B A3B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
Qwen/Qwen3-Next-80B-A3B-Thinking · thinking
 Hugging Face
120.2
From 191 benchmarks. Plenty of data behind it.
90 ±12
Coding IQ 90.0, 95% interval 78–102. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.15$1.50131K
51.1
Fair value (51.1). Costs per task about what its benchmark results predict. Little data behind it, so treat it as a rough estimate.
Not measuredNot measured
Step 3.7 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.
 Llmgateway
120.2
From 41 benchmarks. Plenty of data behind it.
106 ±15
Coding IQ 105.6, 95% interval 90–121. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.20cache read $0.04$1.15Not published by the provider
47.0
Poor value (47.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
57 ±9
Chatty (57 ±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
Step 3.7 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.
stepfun/step-3.7-flash:thinking · thinking · fp8 · 🇺🇸 US
 NanoGPT
120.2
From 41 benchmarks. Plenty of data behind it.
106 ±15
Coding IQ 105.6, 95% interval 90–121. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.21cache read $0.042$1.2075Not published by the provider
47.0
Poor value (47.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
57 ±9
Chatty (57 ±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
Step 3.7 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.
fp8
 OpenRouter
120.2
From 41 benchmarks. Plenty of data behind it.
106 ±15
Coding IQ 105.6, 95% interval 90–121. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.20cache read $0.04$1.15262K
47.0
Poor value (47.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
57 ±9
Chatty (57 ±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
Step 3.7 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.
 Requesty
120.2
From 41 benchmarks. Plenty of data behind it.
106 ±15
Coding IQ 105.6, 95% interval 90–121. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.20cache read $0.04$1.15262K
47.0
Poor value (47.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
57 ±9
Chatty (57 ±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
MiniMax M2.7
🇨🇳
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.
 Llmgateway
119.8
From 207 benchmarks. Plenty of data behind it.
98 ±12
Coding IQ 98.4, 95% interval 87–110. 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$1.20Not published by the provider
50.7
Fair value (50.7). Costs per task about what 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.
Not measured
MiniMax M2.7
🇨🇳
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.
🇺🇸 US
 NanoGPT
119.8
From 207 benchmarks. Plenty of data behind it.
98 ±12
Coding IQ 98.4, 95% interval 87–110. 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
50.7
Fair value (50.7). Costs per task about what 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.
Not measured
MiniMax M2.7
🇨🇳
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.
 Requesty
119.8
From 207 benchmarks. Plenty of data behind it.
98 ±12
Coding IQ 98.4, 95% interval 87–110. 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$1.20200K
50.7
Fair value (50.7). Costs per task about what 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.
Not measured
MiniMax M2.7
🇨🇳
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.
novita/minimax/minimax-m2.7-highspeed · highspeed
 Requesty
119.8
From 207 benchmarks. Plenty of data behind it.
98 ±12
Coding IQ 98.4, 95% interval 87–110. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.60$2.40205K
50.7
Fair value (50.7). Costs per task about what 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.
Not measured
Qwen3.6 35B A3B
🇨🇳
Maker headquartered in CN, according to the company's own website. The citation is on the model page.
Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.
 Llmgateway
119.8
From 160 benchmarks. Plenty of data behind it.
95 ±14
Coding IQ 94.8, 95% interval 81–109. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.248$1.485Not published by the provider
38.5
Poor value (38.5). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
144.1
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API
DeepSeek V3.1
🇨🇳
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.
 Hugging Face
119.6
From 122 benchmarks. Plenty of data behind it.
94 ±8
Coding IQ 94.3, 95% interval 86–103. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.
$0.27$1.00131K
44.0
Poor value (44.0). Costs more per task than its benchmark results predict. Plenty of data behind it.
Not measured
140.9
Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API

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

How this data works