Listings
All listings: Llmgateway
Every listing for the models on page 6 of 13 of Llmgateway, including each model's other versions and routes. Back to Llmgateway
Sold by Llmgateway
84 listings on this page sold by Llmgateway.
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
- Epoch Capabilities Index
- Epoch Capabilities Index benchmark details
| Model | llminfo IQ | Coding IQ | Input $/1M | Output $/1M | Max Input Tokens | Value | Yap | Epoch Capabilities Index |
|---|---|---|---|---|---|---|---|---|
Grok 4.1 Fast 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.9From 260 benchmarks. Plenty of data behind it. | 95 ±14Coding IQ 95.2, 95% interval 82–109. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.20 | $0.50 | 2M | 59.7Good value (59.7). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 56 ±9Chatty (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 |
Grok 4.1 Fast Inference by Microsoft Azure🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.9From 260 benchmarks. Plenty of data behind it. | 95 ±14Coding IQ 95.2, 95% interval 82–109. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.20 | $0.50 | —Not published by the provider | 59.7Good value (59.7). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 56 ±9Chatty (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 |
Grok 4.1 Fast Inference by xAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.9From 260 benchmarks. Plenty of data behind it. | 95 ±14Coding IQ 95.2, 95% interval 82–109. 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 $0.05 | $0.50 | —Not published by the provider | 59.7Good value (59.7). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 56 ±9Chatty (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 |
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. | 122.8From 192 benchmarks. Plenty of data behind it. | 100 ±6Coding 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 | $1.20 | 229K | 48.7Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 64 ±9Chatty (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. | 122.8From 192 benchmarks. Plenty of data behind it. | 100 ±6Coding 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 | $2.40 | 205K | 48.7Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 64 ±9Chatty (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 Inference by MiniMax🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 122.8From 192 benchmarks. Plenty of data behind it. | 100 ±6Coding 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 | $1.20 | —Not published by the provider | 48.7Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 64 ±9Chatty (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 Inference by MiniMax🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 122.8From 192 benchmarks. Plenty of data behind it. | 100 ±6Coding 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 | $2.40 | —Not published by the provider | 48.7Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 64 ±9Chatty (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 Inference by Nebius AI Studio🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 122.8From 192 benchmarks. Plenty of data behind it. | 100 ±6Coding 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. | $0.30 | $1.20 | —Not published by the provider | 48.7Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 64 ±9Chatty (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 Inference by Novita AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 122.8From 192 benchmarks. Plenty of data behind it. | 100 ±6Coding 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 | $1.20 | —Not published by the provider | 48.7Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 64 ±9Chatty (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 Inference by Together AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 122.8From 192 benchmarks. Plenty of data behind it. | 100 ±6Coding 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. | $0.30 | $1.20 | —Not published by the provider | 48.7Fair value (48.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 64 ±9Chatty (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 |
Claude Haiku 4.5 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.6From 470 benchmarks. Plenty of data behind it. | 95 ±6Coding IQ 95.1, 95% interval 90–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $1.00cache read $0.10 | $5.00cache write $1.25 | 200K | 35.0Poor value (35.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±10Chatty (59 ±10). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 142.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Claude Haiku 4.5 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.6From 470 benchmarks. Plenty of data behind it. | 95 ±6Coding IQ 95.1, 95% interval 90–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $1.00cache read $0.10 | $5.00cache write $1.25 | 200K | 35.0Poor value (35.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±10Chatty (59 ±10). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 142.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Claude Haiku 4.5 Inference by Amazon Web Services🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.6From 470 benchmarks. Plenty of data behind it. | 95 ±6Coding IQ 95.1, 95% interval 90–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $1.00cache read $0.10 | $5.00cache write $1.25 | —Not published by the provider | 35.0Poor value (35.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±10Chatty (59 ±10). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 142.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Claude Haiku 4.5 Inference by Amazon Web Services🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.6From 470 benchmarks. Plenty of data behind it. | 95 ±6Coding IQ 95.1, 95% interval 90–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $1.00cache read $0.10 | $5.00cache write $1.25 | —Not published by the provider | 35.0Poor value (35.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±10Chatty (59 ±10). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 142.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Claude Haiku 4.5 Inference by Anthropic🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.6From 470 benchmarks. Plenty of data behind it. | 95 ±6Coding IQ 95.1, 95% interval 90–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $1.00cache read $0.10 | $5.00cache write $1.25 | —Not published by the provider | 35.0Poor value (35.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±10Chatty (59 ±10). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 142.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Claude Haiku 4.5 Inference by Anthropic🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.6From 470 benchmarks. Plenty of data behind it. | 95 ±6Coding IQ 95.1, 95% interval 90–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $1.00cache read $0.10 | $5.00cache write $1.25 | —Not published by the provider | 35.0Poor value (35.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±10Chatty (59 ±10). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 142.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Claude Haiku 4.5 Inference by Google🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | 122.6From 470 benchmarks. Plenty of data behind it. | 95 ±6Coding IQ 95.1, 95% interval 90–101. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $1.00cache read $0.10 | $5.00cache write $1.25 | —Not published by the provider | 35.0Poor value (35.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±10Chatty (59 ±10). Generates more tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 142.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
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. | 122.3From 97 benchmarks. Plenty of data behind it. | 97 ±6Coding 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. | $0.27 | $1.10 | 205K | 45.1Poor value (45.1). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±13Chatty (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 |
MiniMax M2.1 Inference by MiniMax🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 122.3From 97 benchmarks. Plenty of data behind it. | 97 ±6Coding 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. | $0.27 | $1.10 | —Not published by the provider | 45.1Poor value (45.1). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±13Chatty (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 |
MiniMax M2.1 Inference by Novita AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 122.3From 97 benchmarks. Plenty of data behind it. | 97 ±6Coding 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 | $1.20 | —Not published by the provider | 45.1Poor value (45.1). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±13Chatty (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 |
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 | 121.9From 110 benchmarks. Plenty of data behind it. | 95 ±7Coding 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 | $2.20 | 128K | 41.4Poor value (41.4). Costs more per task than its benchmark results predict. Plenty of data behind it. | 53 ±9Typical (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 |
GLM 4.5 Inference by Z.ai🇨🇳Maker headquartered in CN, according to the company's own website. The citation is 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 | 121.9From 110 benchmarks. Plenty of data behind it. | 95 ±7Coding 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 | $2.20 | —Not published by the provider | 41.4Poor value (41.4). Costs more per task than its benchmark results predict. Plenty of data behind it. | 53 ±9Typical (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 |
Qwen3.6 Flash 🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Newer version | 121.9From 108 benchmarks. Plenty of data behind it. | 100 ±12Coding IQ 100.2, 95% interval 88–112. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.25cache read $0.05 | $1.50cache write $0.3125 | 984K | 36.3Poor value (36.3). Costs more per task than its benchmark results predict. A fair amount of data behind it. | 60 ±9Chatty (60 ±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. | 143.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Qwen3.6 Flash Inference by Alibaba Cloud🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Newer version | 121.9From 108 benchmarks. Plenty of data behind it. | 100 ±12Coding IQ 100.2, 95% interval 88–112. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.25cache read $0.05 | $1.50cache write $0.3125 | —Not published by the provider | 36.3Poor value (36.3). Costs more per task than its benchmark results predict. A fair amount of data behind it. | 60 ±9Chatty (60 ±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. | 143.7Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Gemini Embedding 2 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 121.2From 27 benchmarks. Plenty of data behind it. | —Not enough data yet | $0.20 | — | 8K | —Not enough data yet | —Not measured | —Not measured |
Gemini Embedding 2 Inference by Google🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 121.2From 27 benchmarks. Plenty of data behind it. | —Not enough data yet | $0.20 | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not 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 | 121.1From 311 benchmarks. Plenty of data behind it. | 95 ±8Coding 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 | $2.00 | 262K | 40.3Poor value (40.3). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±9Chatty (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. | 144.1Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Qwen3.5 35B A3B Inference by Alibaba Cloud🇨🇳Maker headquartered in CN, according to the company's own website. The citation is 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 | 121.1From 311 benchmarks. Plenty of data behind it. | 95 ±8Coding 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 | $2.00 | —Not published by the provider | 40.3Poor value (40.3). Costs more per task than its benchmark results predict. Plenty of data behind it. | 59 ±9Chatty (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. | 144.1Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
120.7From 486 benchmarks. Plenty of data behind it. | 92 ±6Coding IQ 92.4, 95% interval 87–98. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $2.00cache read $0.50 | $8.00 | —Not published by the provider | 42.8Poor value (42.8). Costs more per task than its benchmark results predict. Plenty of data behind it. | 33 ±8Terse (33 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 136.1Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
120.7From 486 benchmarks. Plenty of data behind it. | 92 ±6Coding IQ 92.4, 95% interval 87–98. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $2.00cache read $0.50 | $8.00 | —Not published by the provider | 42.8Poor value (42.8). Costs more per task than its benchmark results predict. Plenty of data behind it. | 33 ±8Terse (33 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | 136.1Epoch 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. | 120.4From 170 benchmarks. Plenty of data behind it. | ≈ 94 ±18Coding 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. | $0.20cache read $0.11 | $0.88 | 262K | 45.0Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
Qwen3 VL 235B Inference by Alibaba Cloud🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 120.4From 170 benchmarks. Plenty of data behind it. | ≈ 94 ±18Coding 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. | $0.50 | $2.00 | —Not published by the provider | 45.0Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
Qwen3 VL 235B Inference by DeepInfra🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 120.4From 170 benchmarks. Plenty of data behind it. | ≈ 94 ±18Coding 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. | $0.20cache read $0.11 | $0.88 | —Not published by the provider | 45.0Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
Qwen3 VL 235B Inference by Novita AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 120.4From 170 benchmarks. Plenty of data behind it. | ≈ 94 ±18Coding 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 | $1.50 | —Not published by the provider | 45.0Poor value (45.0). Costs more per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
GPT 5.4 Nano Inference by Microsoft Azure🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.3From 252 benchmarks. Plenty of data behind it. | 104 ±6Coding IQ 103.6, 95% interval 97–110. 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 | $1.25 | —Not published by the provider | 49.5Fair value (49.5). Costs per task about what its benchmark results predict. Plenty of data behind it. | 47 ±9Typical (47 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | 146.4Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
GPT 5.4 Nano Inference by OpenAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.3From 252 benchmarks. Plenty of data behind it. | 104 ±6Coding IQ 103.6, 95% interval 97–110. 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 | $1.25 | —Not published by the provider | 49.5Fair value (49.5). Costs per task about what its benchmark results predict. Plenty of data behind it. | 47 ±9Typical (47 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | 146.4Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
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. | 120.2From 41 benchmarks. Plenty of data behind it. | 106 ±15Coding 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 | $1.15 | 262K | 47.0Poor value (47.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 57 ±9Chatty (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 Inference by DeepInfra🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 120.2From 41 benchmarks. Plenty of data behind it. | 106 ±15Coding 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 | $1.15 | —Not published by the provider | 47.0Poor value (47.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 57 ±9Chatty (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 Inference by Novita AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 120.2From 41 benchmarks. Plenty of data behind it. | 106 ±15Coding 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 | $1.15 | —Not published by the provider | 47.0Poor value (47.0). Costs more per task than its benchmark results predict. Plenty of data behind it. | 57 ±9Chatty (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 |
GPT-Image-2.5 Flare 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | ≈ 120From 39 benchmarks. Little data behind it, so treat it as a rough estimate. | —Not enough data yet | $5.00cache read $1.25 | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
GPT-Image-2.5 Flare Inference by Microsoft Azure🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | ≈ 120From 39 benchmarks. Little data behind it, so treat it as a rough estimate. | —Not enough data yet | $5.00cache read $1.25 | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
GPT-Image-2.5 Flare Inference by OpenAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | ≈ 120From 39 benchmarks. Little data behind it, so treat it as a rough estimate. | —Not enough data yet | $5.00cache read $1.25 | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
GPT-Image-2.5 Sunburst 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | ≈ 120From 39 benchmarks. Little data behind it, so treat it as a rough estimate. | —Not enough data yet | $5.00cache read $1.25 | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
GPT-Image-2.5 Sunburst Inference by Microsoft Azure🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | ≈ 120From 39 benchmarks. Little data behind it, so treat it as a rough estimate. | —Not enough data yet | $5.00cache read $1.25 | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
GPT-Image-2.5 Sunburst Inference by OpenAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. Newer version | ≈ 120From 39 benchmarks. Little data behind it, so treat it as a rough estimate. | —Not enough data yet | $5.00cache read $1.25 | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | 2K | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | 2K | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | 2K | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | 2K | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | 2K | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 Inference by xAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 Inference by xAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 Inference by xAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 Inference by xAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
Grok Imagine Image 2.0 Inference by xAI🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 120.0From 17 benchmarks. A fair amount of data behind it. | —Not enough data yet | — | — | —Not published by the provider | —Not enough data yet | —Not measured | —Not measured |
Gemini 3 Pro Image 🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 119.8From 189 benchmarks. A fair amount of data behind it. | —Not enough data yet | $2.00 | $12.00 | 66K | —Not enough data yet | —Not measured | —Not measured |
Gemini 3 Pro Image Inference by Google🇺🇸Maker headquartered in US, according to the company's own website. The citation is on the model page. | 119.8From 189 benchmarks. A fair amount of data behind it. | —Not enough data yet | $2.00 | $12.00 | —Not published by the provider | —Not enough data yet | —Not measured | —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. | 119.8From 207 benchmarks. Plenty of data behind it. | 98 ±12Coding 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 | $1.20 | 205K | 50.7Fair value (50.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 61 ±9Chatty (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. | 119.8From 207 benchmarks. Plenty of data behind it. | 98 ±12Coding 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 | $2.40 | 205K | 50.7Fair value (50.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 61 ±9Chatty (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 Inference by MiniMax🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 119.8From 207 benchmarks. Plenty of data behind it. | 98 ±12Coding 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 | $1.20 | —Not published by the provider | 50.7Fair value (50.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 61 ±9Chatty (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 Inference by MiniMax🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 119.8From 207 benchmarks. Plenty of data behind it. | 98 ±12Coding 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 | $2.40 | —Not published by the provider | 50.7Fair value (50.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 61 ±9Chatty (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 Inference by Novita AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 119.8From 207 benchmarks. Plenty of data behind it. | 98 ±12Coding 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 | $1.20 | —Not published by the provider | 50.7Fair value (50.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 61 ±9Chatty (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 Inference by Together AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 119.8From 207 benchmarks. Plenty of data behind it. | 98 ±12Coding 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 | $1.20 | —Not published by the provider | 50.7Fair value (50.7). Costs per task about what its benchmark results predict. Plenty of data behind it. | 61 ±9Chatty (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. | 119.8From 160 benchmarks. Plenty of data behind it. | 95 ±14Coding 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 | $1.485 | 262K | 38.5Poor value (38.5). Costs more per task than its benchmark results predict. Plenty of data behind it. | —Not measured | 144.1Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Qwen3.6 35B A3B Inference by Alibaba Cloud🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 119.8From 160 benchmarks. Plenty of data behind it. | 95 ±14Coding 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. | $0.375 | $2.25 | —Not published by the provider | 38.5Poor value (38.5). Costs more per task than its benchmark results predict. Plenty of data behind it. | —Not measured | 144.1Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Qwen3.6 35B A3B Inference by Novita AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 119.8From 160 benchmarks. Plenty of data behind it. | 95 ±14Coding 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 | $1.485 | —Not published by the provider | 38.5Poor value (38.5). Costs more per task than its benchmark results predict. Plenty of data behind it. | —Not measured | 144.1Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Devstral 2 123b Instruct 2512 🇫🇷Maker headquartered in FR, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 119.5From 25 benchmarks. Plenty of data behind it. | 97 ±9Coding IQ 97.1, 95% interval 88–106. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.40 | $2.00 | 262K | 50.2Fair value (50.2). Costs per task about what its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
Devstral 2 123b Instruct 2512 Inference by Mistral AI🇫🇷Maker headquartered in FR, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 119.5From 25 benchmarks. Plenty of data behind it. | 97 ±9Coding IQ 97.1, 95% interval 88–106. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.40 | $2.00 | —Not published by the provider | 50.2Fair value (50.2). Costs per task about what its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
MiniMax M2 🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.9From 111 benchmarks. Plenty of data behind it. | 95 ±7Coding IQ 95.4, 95% interval 88–103. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.20cache read $0.03 | $1.00 | 197K | 57.8Good value (57.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
MiniMax M2 Inference by MiniMax🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.9From 111 benchmarks. Plenty of data behind it. | 95 ±7Coding IQ 95.4, 95% interval 88–103. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.20cache read $0.03 | $1.00 | —Not published by the provider | 57.8Good value (57.8). Cheaper per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
Seed 2.0 Miniseed-2-0-mini-260428 · 260428seed-2-0-mini-260428 · 260428 | 118.6From 76 benchmarks. Plenty of data behind it. | —Not enough data yet | $0.10cache read $0.02 | $0.40 | 262K | 45.0Poor value (45.0). Costs more per task than its benchmark results predict. Little data behind it, so treat it as a rough estimate. | —Not measured | —Not measured |
GLM 4.6v 🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.4From 45 benchmarks. A fair amount of data behind it. | —Not enough data yet | $0.30cache read $0.05 | $0.90 | 131K | 44.9Poor value (44.9). Costs more per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
GLM 4.6v Inference by Novita AI🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.4From 45 benchmarks. A fair amount of data behind it. | —Not enough data yet | $0.30cache read $0.055 | $0.90 | —Not published by the provider | 44.9Poor value (44.9). Costs more per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
GLM 4.6v Inference by Z.ai🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.4From 45 benchmarks. A fair amount of data behind it. | —Not enough data yet | $0.30cache read $0.05 | $0.90 | —Not published by the provider | 44.9Poor value (44.9). Costs more per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured |
GPT OSS 120B 🇺🇸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. | 118.2From 424 benchmarks. Plenty of data behind it. | 86 ±5Coding IQ 85.5, 95% interval 80–91. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.03cache read $0.0045 | $0.15 | 131K | 58.4Good value (58.4). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 37 ±11Terse (37 ±11). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | 140.2Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
GPT OSS 120B Inference by Microsoft Azure🇺🇸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. | 118.2From 424 benchmarks. Plenty of data behind it. | 86 ±5Coding IQ 85.5, 95% interval 80–91. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.15 | $0.60 | —Not published by the provider | 58.4Good value (58.4). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 37 ±11Terse (37 ±11). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | 140.2Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
GPT OSS 120B Inference by Cerebras🇺🇸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. | 118.2From 424 benchmarks. Plenty of data behind it. | 86 ±5Coding IQ 85.5, 95% interval 80–91. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.35 | $0.75 | —Not published by the provider | 58.4Good value (58.4). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 37 ±11Terse (37 ±11). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | 140.2Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
GPT OSS 120B Inference by Groq🇺🇸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. | 118.2From 424 benchmarks. Plenty of data behind it. | 86 ±5Coding IQ 85.5, 95% interval 80–91. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.15 | $0.75 | —Not published by the provider | 58.4Good value (58.4). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 37 ±11Terse (37 ±11). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | 140.2Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
GPT OSS 120B Inference by Nebius AI Studio🇺🇸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. | 118.2From 424 benchmarks. Plenty of data behind it. | 86 ±5Coding IQ 85.5, 95% interval 80–91. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.15 | $0.60 | —Not published by the provider | 58.4Good value (58.4). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 37 ±11Terse (37 ±11). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | 140.2Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
GPT OSS 120B Inference by Together AI🇺🇸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. | 118.2From 424 benchmarks. Plenty of data behind it. | 86 ±5Coding IQ 85.5, 95% interval 80–91. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.15 | $0.60 | —Not published by the provider | 58.4Good value (58.4). Cheaper per task than its benchmark results predict. Plenty of data behind it. | 37 ±11Terse (37 ±11). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | 140.2Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API |
Mistral Large 3 🇫🇷Maker headquartered in FR, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.1From 60 benchmarks. Plenty of data behind it. | 86 ±8Coding IQ 86.4, 95% interval 79–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.50 | $1.50 | 262K | 38.5Poor value (38.5). Costs more per task than its benchmark results predict. Plenty of data behind it. | 40 ±9Terse (40 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | —Not measured |
Mistral Large 3 Inference by Mistral AI🇫🇷Maker headquartered in FR, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.1From 60 benchmarks. Plenty of data behind it. | 86 ±8Coding IQ 86.4, 95% interval 79–94. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.50 | $1.50 | —Not published by the provider | 38.5Poor value (38.5). Costs more per task than its benchmark results predict. Plenty of data behind it. | 40 ±9Terse (40 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | —Not measured |
GLM 4.5 Air 🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.0From 210 benchmarks. Plenty of data behind it. | 93 ±7Coding IQ 93.2, 95% interval 86–100. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.20cache read $0.03 | $1.10 | 128K | 42.6Poor value (42.6). Costs more per task than its benchmark results predict. Plenty of data behind it. | —Not measured | —Not measured |
GLM 4.5 Air Inference by Z.ai🇨🇳Maker headquartered in CN, according to the company's own website. The citation is on the model page. Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo. | 118.0From 210 benchmarks. Plenty of data behind it. | 93 ±7Coding IQ 93.2, 95% interval 86–100. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | $0.20cache read $0.03 | $1.10 | —Not published by the provider | 42.6Poor value (42.6). Costs more per task than its benchmark results predict. Plenty of data behind it. | —Not measured | —Not measured |
Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence)