Dtbench — Epoch AI
Dtbench — Epoch AI is run by Epoch AI. It has ranked 235 models, of which the catalogue holds 152, scoring from 41.65 to 98.93.
measured by Epoch AI · the board itself
| Place | Model | Metric | Score |
|---|---|---|---|
| 1stof 235 | Claude Opus 5.5 | Accuracy (max) | 98.93 |
| 2ndof 235 | Claude Fable 5 | Accuracy (max) | 98.4 |
| 3rdof 235 | Claude Opus 5 | Accuracy (xhigh) | 97.87 |
| 6thof 235 | Claude Fable 5.1 | Accuracy | 97.6 |
| 8thof 235 | Gemini 4 Argon | Accuracy (low) | 97.6 |
| 9thof 235 | Grok 4.6 | Accuracy (xhigh) | 97.33 |
| 10thof 235 | GPT 6 Astra | Accuracy (xhigh) | 97.33 |
| 12thof 235 | GPT 6 Sol | Accuracy (max) | 97.33 |
| 13thof 235 | Claude Sonnet 5.5 | Accuracy (max) | 97.33 |
| 20thof 235 | Gemini 3.1 Pro Preview | Accuracy (medium) | 97.07 |
| 21stof 235 | Gemini 3.7 Flash | Accuracy (high) | 96.8 |
| 22ndof 235 | Grok 4.5 | Accuracy (high) | 96.53 |
| 26thof 235 | Muse Spark 1.3 | Accuracy | 96.53 |
| 29thof 235 | GPT 6.1 Sol | Accuracy (max) | 96.53 |
| 31stof 235 | GPT-5.5 | Accuracy (xhigh) | 96 |
| 32ndof 235 | GPT-5.5 Pro | Accuracy (xhigh) | 96 |
| 33rdof 235 | Grok 4.7 | Accuracy (xhigh) | 96 |
| 36thof 235 | GPT-5.6 Sol | Accuracy (xhigh) | 95.73 |
| 37thof 235 | Gemini 3.8 Flash | Accuracy | 95.72 |
| 40thof 235 | Gemini 3.6 Flash | Accuracy (high) | 95.47 |
| 45thof 235 | Claude Haiku 5.5 | Accuracy (max) | 94.99 |
| 46thof 235 | Claude Opus 4.8 | Accuracy (max) | 94.93 |
| 52ndof 235 | Muse Spark 1.2 | Accuracy (xhigh) | 94.67 |
| 53rdof 235 | Gemini 3.5 Flash | Accuracy (high) | 94.67 |
| 55thof 235 | Claude Opus 4.7 | Accuracy (max) | 94.65 |
| 56thof 235 | GPT-5.4 | Accuracy (xhigh) | 94.4 |
| 57thof 235 | Muse Spark 1.1 | Accuracy (high) | 94.4 |
| 61stof 235 | DeepSeek V4 Pro 0813 | Accuracy | 93.87 |
| 63rdof 235 | GLM 5.2 | Accuracy (max) | 93.6 |
| 64thof 235 | GPT-5.6 Terra | Accuracy (max) | 93.33 |
| 69thof 235 | Claude Sonnet 5 | Accuracy (max) | 92.48 |
| 70thof 235 | Qwen3.7 Max | Accuracy (max) | 92.27 |
| 71stof 235 | Qwen 3.8 Max | Accuracy (xhigh) | 92 |
| 73rdof 235 | Claude Opus 4.6 | Accuracy (max) | 91.2 |
| 74thof 235 | Kimi K3 | Accuracy (max) | 91.2 |
| 76thof 235 | GPT-5.2 | Accuracy (xhigh) | 90.93 |
| 77thof 235 | DeepSeek V4 Flash (0731) | Accuracy (max) | 90.93 |
| 78thof 235 | Kimi K2.6 | Accuracy | 90.93 |
| 79thof 235 | DeepSeek V4 Pro | Accuracy (max) | 90.67 |
| 80thof 235 | GPT-5 | Accuracy (high) | 90.67 |
| 81stof 235 | Grok 4.3 | Accuracy (high) | 90.67 |
| 82ndof 235 | GPT-5.1 | Accuracy (high) | 90.13 |
| 83rdof 235 | Grok 4.20 | Accuracy | 90.13 |
| 84thof 235 | Nemotron 3 Ultra | Accuracy | 90.13 |
| 87thof 235 | GPT 6 Luna | Accuracy (max) | 90.13 |
| 88thof 235 | Claude Sonnet 4.6 | Accuracy (max) | 89.87 |
| 89thof 235 | Claude Opus 4.5 | Accuracy (high) | 89.87 |
| 90thof 235 | DeepSeek V4.1 Flash | Accuracy | 89.87 |
| 92ndof 235 | Gemini 3 Flash | Accuracy (high) | 89.07 |
| 93rdof 235 | GPT-5.6 Luna | Accuracy (xhigh) | 89.07 |
| 96thof 235 | Qwen3.8 27B | Accuracy | 88 |
| 97thof 235 | Grok 4.1 Fast Reasoning | Accuracy | 87.73 |
| 99thof 235 | GLM-5.3 | Accuracy | 87.67 |
| 100thof 235 | Inkling | Accuracy (xhigh) | 87.47 |
| 101stof 235 | Qwen3.5 397B A17B | Accuracy | 87.47 |
| 102ndof 235 | Qwen3.6 Max Preview | Accuracy (max) | 87.2 |
| 104thof 235 | o3-pro | Accuracy (high) | 86.93 |
| 107thof 235 | DeepSeek V4 Flash | Accuracy (max) | 86.4 |
| 108thof 235 | DeepSeek v3.2 | Accuracy | 85.6 |
| 109thof 235 | o3 | Accuracy (high) | 84.8 |
| 110thof 235 | MiMo-V2.5-Pro | Accuracy | 84.53 |
| 112thof 235 | Qwen3.5-122B-A10B | Accuracy (none) | 84.27 |
| 114thof 235 | Qwen 3.7 Plus | Accuracy | 84 |
| 116thof 235 | Gemini 3.5 Flash-Lite | Accuracy (high) | 83.47 |
| 117thof 235 | Claude Sonnet 4.5 | Accuracy | 83.2 |
| 119thof 235 | Qwen3.5-Flash | Accuracy | 82.93 |
| 120thof 235 | Grok 4 Fast | Accuracy | 82.67 |
| 121stof 235 | Gemma 4 31B | Accuracy | 82.67 |
| 123rdof 235 | Gemini 2.5 Pro | Accuracy | 82.4 |
| 124thof 235 | Qwen3.5-27B | Accuracy | 82.4 |
| 126thof 235 | Qwen3.6 Plus | Accuracy | 81.87 |
| 127thof 235 | Claude Opus 4 | Accuracy | 81.6 |
| 128thof 235 | DeepSeek V3.1 Terminus | Accuracy | 81.33 |
| 129thof 235 | Qwen-Plus | Accuracy | 81.07 |
| 131stof 235 | GPT-5 mini | Accuracy (high) | 80.53 |
| 132ndof 235 | Qwen 3.5 Plus | Accuracy | 80.53 |
| 133rdof 235 | GPT-5.4 nano | Accuracy (xhigh) | 80.27 |
| 134thof 235 | Qwen3 235B A22B Thinking 2507 | Accuracy | 80.27 |
| 135thof 235 | GPT-5.4 mini | Accuracy (xhigh) | 80 |
| 136thof 235 | Claude Opus 4.1 | Accuracy | 80 |
| 137thof 235 | Qwen3.5-35B-A3B | Accuracy | 80 |
| 140thof 235 | MiniMax M3 | Accuracy | 78.93 |
| 142ndof 235 | Qwen3-235B-A22B-Instruct-2507 | Accuracy | 78.4 |
| 143rdof 235 | Qwen3.6 27B | Accuracy | 78.13 |
| 144thof 235 | o4-mini | Accuracy (high) | 77.6 |
| 145thof 235 | Qwen3.6-Flash | Accuracy | 77.07 |
| 146thof 235 | Claude Sonnet 4 | Accuracy | 77.07 |
| 147thof 235 | Gemini 3.1 Flash-Lite | Accuracy (high) | 76.8 |
| 148thof 235 | Gemini 2.5 Flash | Accuracy | 76.53 |
| 149thof 235 | GPT OSS 120B | Accuracy (high) | 76.27 |
| 150thof 235 | Qwen3 235B A22B | Accuracy | 75.73 |
| 153rdof 235 | Gemma 4 26B A4B | Accuracy | 74.93 |
| 154thof 235 | o1 | Accuracy (high) | 74.67 |
| 155thof 235 | Qwen3.6 35B A3B | Accuracy | 73.87 |
| 156thof 235 | Claude Haiku 4.5 | Accuracy | 73.6 |
| 158thof 235 | Qwen3.5 9B | Accuracy | 71.2 |
| 159thof 235 | Mistral Small 4 | Accuracy | 70.93 |
| 160thof 235 | Qwen3 30B A3B Thinking 2507 | Accuracy | 69.33 |
| 162ndof 235 | o3-mini | Accuracy (high) | 68.8 |
| 163rdof 235 | GPT-4.1 mini | Accuracy | 68.8 |
| 164thof 235 | GPT-4.1 fine-tuned | Accuracy | 68.27 |
| 165thof 235 | GPT OSS 20B | Accuracy (high) | 68 |
| 166thof 235 | Claude 3.5 Sonnet | Accuracy | 67.84 |
| 168thof 235 | Qwen3 32B | Accuracy | 67.47 |
| 170thof 235 | Qwen3 30B A3B Instruct 2507 | Accuracy | 67.2 |
| 172ndof 235 | Mistral Large 3 | Accuracy | 65.07 |
| 173rdof 235 | DeepSeek-V3 0324 | Accuracy | 64.8 |
| 174thof 235 | GPT-4o (2024-05-13) | Accuracy | 64.53 |
| 175thof 235 | Qwen3 14B | Accuracy | 64 |
| 176thof 235 | gemini-2.0-flash-001 | Accuracy | 63.2 |
| 177thof 235 | Qwen2.5 72B Instruct | Accuracy | 62.93 |
| 179thof 235 | GPT-4 (0613) | Accuracy | 62.67 |
| 180thof 235 | deepseek-chat | Accuracy | 62.67 |
| 181stof 235 | GPT-5 nano | Accuracy (high) | 62.67 |
| 182ndof 235 | Mistral Medium 2505 | Accuracy | 62.27 |
| 183rdof 235 | Llama-4-Maverick-17B-128E-Instruct | Accuracy | 61.87 |
| 184thof 235 | Claude 3 Opus | Accuracy | 61.6 |
| 185thof 235 | GPT-4 Turbo | Accuracy | 61.6 |
| 186thof 235 | Meta-Llama-3.1-405B-Instruct | Accuracy | 61.38 |
| 187thof 235 | c4ai-command-a-03-2025 | Accuracy | 61.33 |
| 189thof 235 | Mistral Large 2407 | Accuracy | 61.21 |
| 191stof 235 | Qwen3 30B A3B | Accuracy | 60.27 |
| 192ndof 235 | Meta-Llama-3.1-70B-Instruct | Accuracy | 60 |
| 194thof 235 | Qwen3 8B | Accuracy | 59.73 |
| 195thof 235 | Llama 3.3 70B Instruct | Accuracy | 59.47 |
| 197thof 235 | mistral-small-2503 | Accuracy | 58.64 |
| 199thof 235 | Claude Haiku 3.5 | Accuracy | 56.66 |
| 201stof 235 | Mistral Large (24.02) | Accuracy | 55.9 |
| 202ndof 235 | Mixtral-8x22B-Instruct-v0.1 | Accuracy | 55.14 |
| 203rdof 235 | c4ai-command-r-plus-08-2024 | Accuracy | 54.93 |
| 204thof 235 | GPT-4o-mini (2024-07-18) | Accuracy | 54.4 |
| 205thof 235 | Meta-Llama-3-70B-Instruct | Accuracy | 54.16 |
| 207thof 235 | Claude 3 Sonnet | Accuracy | 53.6 |
| 209thof 235 | Mistral Small (24.02) | Accuracy | 52.96 |
| 210thof 235 | Gemini 2.0 Flash-Lite | Accuracy | 52.53 |
| 211thof 235 | Gemma 3 27B | Accuracy | 52.53 |
| 212thof 235 | GPT-4.1 nano | Accuracy | 52.53 |
| 213thof 235 | Claude 2.0 | Accuracy | 51.85 |
| 216thof 235 | Claude 2.1 | Accuracy | 50.95 |
| 217thof 235 | Llama 3.1 8B Instruct | Accuracy | 50.93 |
| 218thof 235 | Gemma 3 4B | Accuracy | 50.93 |
| 219thof 235 | Claude 3 Haiku | Accuracy | 50.13 |
| 223rdof 235 | Gemma 3 12B | Accuracy | 48.8 |
| 224thof 235 | Mistral Nemo | Accuracy | 48.61 |
| 225thof 235 | GPT-3.5 Turbo | Accuracy | 48.53 |
| 226thof 235 | Gemma 2 27B | Accuracy | 48 |
| 227thof 235 | Qwen2.5 7B Instruct | Accuracy | 47.73 |
| 228thof 235 | c4ai-command-r-08-2024 | Accuracy | 46.4 |
| 232ndof 235 | Llama 3 8B Instruct | Accuracy | 43.93 |
| 233rdof 235 | open-mistral-7b | Accuracy | 42.54 |
| 234thof 235 | Llama-2-13b-chat | Accuracy | 42.24 |
| 235thof 235 | Llama-2-70b-chat | Accuracy | 41.65 |
Read from the board on 2026-10-31