Knowledge & reasoning
SuperGPQA leaderboard
As of 2026-11-21, the highest measured score on SuperGPQA is 95.0% by Claude Opus 4.6. 261 more models have estimated scores, calibrated from the benchmarks they were measured on.
Measured scores: benchlm.ai.
| # | Model | Score | Source |
|---|---|---|---|
| 1 | Claude Fable 5 | 100.0% | estimated ± 9.3 pp, low confidence |
| 2 | Claude Fable 5.1 | 100.0% | estimated ± 7.2 pp, low confidence |
| 3 | Claude Fable 5.2 | 100.0% | estimated ± 7.2 pp, low confidence |
| 4 | Claude Mythos 5 | 100.0% | estimated ± 7.2 pp, low confidence |
| 5 | Claude Opus 5 | 100.0% | estimated ± 7.2 pp, low confidence |
| 6 | Claude Opus 5.5 | 100.0% | estimated ± 9.3 pp, low confidence |
| 7 | Claude Sonnet 5.5 | 100.0% | estimated ± 9.3 pp, low confidence |
| 8 | Gemini 3.7 Flash | 100.0% | estimated ± 9.3 pp, low confidence |
| 9 | Gemini 3 Pro | 100.0% | estimated ± 9.3 pp, low confidence |
| 10 | Gemini 4 Carbon | 100.0% | estimated ± 9.3 pp, low confidence |
| 11 | Gemini 4 Neon | 100.0% | estimated ± 9.3 pp, low confidence |
| 12 | GPT-6.1 Sol | 100.0% | estimated ± 9.3 pp, low confidence |
| 13 | GPT-6 Sol | 100.0% | estimated ± 9.3 pp, low confidence |
| 14 | Grok 4.5 | 100.0% | estimated ± 9.3 pp, low confidence |
| 15 | Kimi K3.1 Pro | 100.0% | estimated ± 7.2 pp, low confidence |
| 16 | Muse Spark 1.1 | 100.0% | estimated ± 7.2 pp, low confidence |
| 17 | Gemini 3.8 Flash | 99.9% | estimated ± 9.3 pp, low confidence |
| 18 | Gemini 4 Hydrogen | 99.9% | estimated ± 9.3 pp, low confidence |
| 19 | Gemini 3.6 Flash | 99.1% | estimated ± 9.3 pp, low confidence |
| 20 | Hy4 | 97.8% | estimated ± 7.2 pp, low confidence |
| 21 | GPT-5.4 Pro | 97.4% | estimated ± 7.2 pp, low confidence |
| 22 | Claude Opus 4.8 | 96.7% | estimated ± 7.2 pp, low confidence |
| 23 | Gemini 4 Argon | 96.3% | estimated ± 9.3 pp, low confidence |
| 24 | Claude Sonnet 5 | 96.2% | estimated ± 7.2 pp, low confidence |
| 25 | Grok 4.6 | 96.2% | estimated ± 9.3 pp, low confidence |
| 26 | GPT-5.5 Pro | 96.0% | estimated ± 7.2 pp, low confidence |
| 27 | GPT-6.1 Terra | 95.1% | estimated ± 9.3 pp, low confidence |
| 28 | Claude Opus 4.6 | 95.0% | measured |
| 29 | Claude Sonnet 4.6 | 95.0% | measured |
| 30 | Apodex 1.1 | 95.0% | estimated ± 7.2 pp, low confidence |
| 31 | Kimi K3 | 94.9% | estimated ± 7.2 pp, low confidence |
| 32 | Hy4 preview | 94.3% | estimated ± 7.2 pp, low confidence |
| 33 | Hy4 flash | 94.1% | estimated ± 7.2 pp, low confidence |
| 34 | Claude Opus 4.6 (Adaptive) | 94.1% | estimated ± 9.3 pp, low confidence |
| 35 | Claude Opus 4.7 | 93.7% | estimated ± 7.2 pp, low confidence |
| 36 | Claude Opus 4.5 Thinking | 93.4% | estimated ± 9.3 pp, low confidence |
| 37 | Inkling 2 Small | 92.3% | estimated ± 7.2 pp, low confidence |
| 38 | Gemini 3 Flash | 91.9% | estimated ± 9.3 pp, medium confidence |
| 39 | dots3-note Preview | 91.7% | estimated ± 7.2 pp, medium confidence |
| 40 | GPT-5.5 | 91.3% | estimated ± 7.2 pp, medium confidence |
| 41 | GPT-5.4 | 91.2% | estimated ± 7.2 pp, medium confidence |
| 42 | Muse Spark 1.2 | 91.2% | estimated ± 9.3 pp, medium confidence |
| 43 | Qwen4 Max | 90.8% | estimated ± 7.2 pp, medium confidence |
| 44 | MiniMax M3.1 Flash | 90.6% | estimated ± 9.3 pp, medium confidence |
| 45 | MiniMax M3 | 90.3% | estimated ± 9.3 pp, medium confidence |
| 46 | Muse Spark | 89.6% | estimated ± 7.2 pp, medium confidence |
| 47 | Muse Spark 1.3 | 87.8% | estimated ± 9.3 pp, medium confidence |
| 48 | MiMo-V2.5-Pro | 87.4% | estimated ± 7.2 pp, medium confidence |
| 49 | Inkling-Small | 87.2% | estimated ± 7.2 pp, medium confidence |
| 50 | Agents-A1 | 87.0% | estimated ± 7.2 pp, medium confidence |
| 51 | Grok 4.7 | 86.2% | estimated ± 9.3 pp, medium confidence |
| 52 | Step 5 Preview | 86.0% | estimated ± 7.2 pp, medium confidence |
| 53 | Inkling | 85.5% | estimated ± 7.2 pp, medium confidence |
| 54 | Muse Spark Open | 84.9% | estimated ± 9.3 pp, medium confidence |
| 55 | GPT-6.2 Astra | 84.7% | estimated ± 8.5 pp, low confidence |
| 56 | Ornith-1.5-397B | 84.2% | estimated ± 7.2 pp, medium confidence |
| 57 | GPT-6.2 Sol | 83.4% | estimated ± 8.5 pp, low confidence |
| 58 | GPT-6 Astra | 83.4% | estimated ± 8.5 pp, low confidence |
| 59 | Qwen3.8 Max | 83.2% | estimated ± 7.2 pp, medium confidence |
| 60 | GPT-5.2-Codex | 83.2% | estimated ± 9.3 pp, medium confidence |
| 61 | Gemini 3.1 Pro | 83.0% | estimated ± 13.9 pp, low confidence |
| 62 | Muse Wonder Full | 83.0% | estimated ± 9.3 pp, medium confidence |
| 63 | Sakana Fugu | 82.8% | estimated ± 8.5 pp, low confidence |
| 64 | Sakana Fugu-Ultra | 82.8% | estimated ± 8.5 pp, low confidence |
| 65 | DeepSeek V4 Pro | 82.4% | estimated ± 7.2 pp, medium confidence |
| 66 | DeepSeek V4.1 Flash 1017 | 82.3% | estimated ± 7.2 pp, medium confidence |
| 67 | Grok 4 | 82.1% | estimated ± 9.3 pp, medium confidence |
| 68 | GPT-5 | 81.8% | estimated ± 9.3 pp, medium confidence |
| 69 | GPT-5.6 Sol | 81.6% | estimated ± 8.5 pp, low confidence |
| 70 | GPT-5.4 mini | 81.3% | estimated ± 7.2 pp, medium confidence |
| 71 | GPT-5.1-Codex | 81.1% | estimated ± 9.3 pp, medium confidence |
| 72 | GPT-5.1-Codex-Max | 81.1% | estimated ± 9.3 pp, medium confidence |
| 73 | GPT-6 Luna | 80.5% | estimated ± 9.3 pp, medium confidence |
| 74 | GLM-5.2 | 80.3% | estimated ± 7.2 pp, medium confidence |
| 75 | Pareto 26.10 Preview | 80.2% | estimated ± 13.9 pp, low confidence |
| 76 | Gemini 3.5 Flash | 80.0% | estimated ± 7.2 pp, medium confidence |
| 77 | GPT-5.6 Terra | 79.3% | estimated ± 8.5 pp, low confidence |
| 78 | GPT-5.2 | 78.6% | estimated ± 8.5 pp, low confidence |
| 79 | GPT-5.6 Luna | 78.5% | estimated ± 8.5 pp, low confidence |
| 80 | Muse Wonder | 77.8% | estimated ± 9.3 pp, medium confidence |
| 81 | GPT-5.4 nano | 77.7% | estimated ± 7.2 pp, medium confidence |
| 82 | Ling 3.0 Tiny | 77.7% | estimated ± 9.3 pp, medium confidence |
| 83 | Ling 3.1 Flash | 77.5% | estimated ± 7.2 pp, medium confidence |
| 84 | GPT-5.1 | 77.4% | estimated ± 9.3 pp, medium confidence |
| 85 | Particle-1 | 77.0% | estimated ± 13.9 pp, low confidence |
| 86 | Beam 1.0 501B-A23B FP8 | 76.9% | estimated ± 13.9 pp, low confidence |
| 87 | Qwen3.8-Omni-Flash | 76.6% | estimated ± 7.2 pp, medium confidence |
| 88 | Qwen3.8-Flash-Next | 76.0% | estimated ± 7.2 pp, medium confidence |
| 89 | Beam 1.0 501B-A23B NVFP4 | 75.3% | estimated ± 13.9 pp, low confidence |
| 90 | Grok 4.3 | 75.2% | estimated ± 7.2 pp, medium confidence |
| 91 | Interfaze Beta | 75.2% | estimated ± 8.5 pp, low confidence |
| 92 | Mercury 3 | 75.1% | estimated ± 13.9 pp, low confidence |
| 93 | DeepSeek V4 Flash 0731 | 75.0% | estimated ± 7.2 pp, medium confidence |
| 94 | Kimi K2.6 | 74.9% | estimated ± 7.2 pp, medium confidence |
| 95 | Grok 4.20 | 73.9% | estimated ± 13.9 pp, low confidence |
| 96 | Qwen 3.6 Max (preview) | 73.9% | measured |
| 97 | Qwen3.7 Max | 73.6% | measured |
| 98 | OrcaCyber-Nano-1.0 | 73.6% | estimated ± 13.9 pp, low confidence |
| 99 | MiMo-V2.6-Pro | 73.2% | estimated ± 9.3 pp, medium confidence |
| 100 | Claude Haiku 5.5 | 73.1% | estimated ± 9.3 pp, medium confidence |
| 101 | Qwen3.6 Plus | 71.6% | measured |
| 102 | GLM-5.1 | 71.4% | estimated ± 7.2 pp, medium confidence |
| 103 | Qwen3.7 Plus | 71.4% | measured |
| 104 | Qwen3.8-27B | 71.2% | estimated ± 7.2 pp, medium confidence |
| 105 | Seed 2.1 Pro | 70.8% | measured |
| 106 | Claude Opus 4.5 | 70.6% | measured |
| 107 | Inkling 2 Mini | 70.5% | estimated ± 8.5 pp, low confidence |
| 108 | Nemotron 3.5 Bolt 74B-A9B | 70.5% | estimated ± 13.9 pp, low confidence |
| 109 | Qwen3.5 397B | 70.4% | measured |
| 110 | GLM-5.3 | 70.3% | estimated ± 9.3 pp, medium confidence |
| 111 | Solar Open 2 | 70.2% | estimated ± 12.4 pp, low confidence |
| 112 | Nemotron 3.5 Super VL 120B A12B | 69.7% | estimated ± 7.2 pp, medium confidence |
| 113 | Nemotron 3.5 Super VL 120B A12B NVFP4 | 69.5% | estimated ± 7.2 pp, medium confidence |
| 114 | Kimi K2.5 | 69.2% | measured |
| 115 | Beam 1.0 501B-A23B | 68.9% | estimated ± 9.3 pp, medium confidence |
| 116 | MiniMax M2.7 | 68.6% | estimated ± 7.2 pp, medium confidence |
| 117 | A.X K2 | 68.4% | estimated ± 7.2 pp, medium confidence |
| 118 | Qwen3.8 Max Preview | 68.0% | estimated ± 9.3 pp, medium confidence |
| 119 | GLM-5V-Turbo | 67.9% | estimated ± 9.3 pp, medium confidence |
| 120 | Seed 2.1 Turbo | 67.4% | measured |
| 121 | Nemotron 3 Ultra | 67.4% | estimated ± 7.2 pp, medium confidence |
| 122 | Ripple-65B Preview | 67.4% | estimated ± 12.4 pp, low confidence |
| 123 | GLM-5-Turbo | 67.3% | estimated ± 9.3 pp, medium confidence |
| 124 | Gemma 4 31B | 67.2% | estimated ± 7.2 pp, medium confidence |
| 125 | Apodex 1.1 Mini | 67.2% | estimated ± 9.3 pp, medium confidence |
| 126 | Qwen3.5-122B-A10B | 67.1% | measured |
| 127 | MAI-Thinking-1 | 67.1% | estimated ± 8.5 pp, low confidence |
| 128 | Muse Glimmer 30B | 67.1% | estimated ± 9.3 pp, medium confidence |
| 129 | Ling 3.0 Flash FP8 | 66.8% | estimated ± 8.5 pp, low confidence |
| 130 | GLM-5 | 66.8% | measured |
| 131 | DeepSeek-R1-0528 | 66.7% | estimated ± 4.1 pp, high confidence |
| 132 | MiMo-V2.6-Flash | 66.6% | estimated ± 9.3 pp, medium confidence |
| 133 | MiMo-V2-Flash | 66.4% | estimated ± 8.5 pp, low confidence |
| 134 | MiMo-V2-Pro | 66.4% | estimated ± 9.3 pp, medium confidence |
| 135 | Ornith-1.5-35B-A3B | 66.3% | estimated ± 7.2 pp, medium confidence |
| 136 | Gemini 3.5 Flash-Lite | 66.3% | estimated ± 9.3 pp, medium confidence |
| 137 | Hy3 Preview | 66.2% | estimated ± 7.2 pp, medium confidence |
| 138 | Gemini 2.5 Flash | 66.2% | estimated ± 9.3 pp, medium confidence |
| 139 | Mistral Large 4 | 66.1% | estimated ± 9.3 pp, medium confidence |
| 140 | Step 3.7 Flash | 66.0% | estimated ± 9.3 pp, medium confidence |
| 141 | Qwen3.6-27B | 66.0% | measured |
| 142 | Claude Sonnet 4.5 | 66.0% | estimated ± 8.5 pp, low confidence |
| 143 | Grok 4.1 Fast (Reasoning) | 65.8% | estimated ± 9.3 pp, medium confidence |
| 144 | Mistral Large 3 | 65.8% | estimated ± 9.3 pp, medium confidence |
| 145 | Qwen3.5-27B | 65.6% | measured |
| 146 | GLM-4.7 | 65.6% | estimated ± 7.2 pp, medium confidence |
| 147 | Qwen3 Max | 65.6% | estimated ± 9.3 pp, medium confidence |
| 148 | DeepSeek V3 0324 | 65.5% | estimated ± 9.3 pp, medium confidence |
| 149 | DeepSeek V3.2 | 65.5% | estimated ± 9.3 pp, medium confidence |
| 150 | Grok Code Fast 1 | 65.3% | estimated ± 9.3 pp, medium confidence |
| 151 | Mistral Medium 3.5 128B | 65.2% | estimated ± 9.3 pp, medium confidence |
| 152 | Grok 4 Fast (Reasoning) | 65.2% | estimated ± 9.3 pp, medium confidence |
| 153 | Claude 4 Sonnet | 65.1% | estimated ± 9.3 pp, medium confidence |
| 154 | Mercury 2.5 | 65.0% | estimated ± 9.3 pp, medium confidence |
| 155 | Nemotron 3.5 Bolt 74B-A9B BF16 | 65.0% | estimated ± 9.3 pp, medium confidence |
| 156 | GLM-4.6 | 64.9% | estimated ± 9.3 pp, medium confidence |
| 157 | Mistral Small 4 | 64.9% | estimated ± 9.3 pp, medium confidence |
| 158 | Muse Ember | 64.9% | estimated ± 9.3 pp, medium confidence |
| 159 | GPT-4o | 64.7% | estimated ± 9.3 pp, medium confidence |
| 160 | Mistral Large 2 | 64.7% | estimated ± 9.3 pp, medium confidence |
| 161 | Qwen3.6-35B-A3B | 64.7% | measured |
| 162 | MiMo-V2-Omni | 64.7% | estimated ± 9.3 pp, low confidence |
| 163 | Solar Pro 3 | 64.6% | estimated ± 9.3 pp, low confidence |
| 164 | North Mini Code | 64.6% | estimated ± 9.3 pp, low confidence |
| 165 | Solar Pro 4 | 64.6% | estimated ± 9.3 pp, low confidence |
| 166 | Mistral Medium 3 | 64.6% | estimated ± 9.3 pp, low confidence |
| 167 | Claude 3 Haiku | 64.5% | estimated ± 9.3 pp, low confidence |
| 168 | Grok 4.1 Fast | 64.5% | estimated ± 9.3 pp, low confidence |
| 169 | Nova Pro | 64.5% | estimated ± 9.3 pp, low confidence |
| 170 | K-Exaone | 64.5% | estimated ± 9.3 pp, low confidence |
| 171 | GLM-4.5-Air | 64.5% | estimated ± 9.3 pp, low confidence |
| 172 | Solar Pro 2 | 64.4% | estimated ± 9.3 pp, low confidence |
| 173 | Qwen3 235B-A22B (Reasoning) | 64.4% | estimated ± 2.0 pp, low confidence |
| 174 | Mistral Small 3.1 | 64.4% | estimated ± 9.3 pp, low confidence |
| 175 | Devstral Small | 64.4% | estimated ± 9.3 pp, low confidence |
| 176 | GPT-4.1 nano | 64.4% | estimated ± 9.3 pp, low confidence |
| 177 | Claude 3.5 Haiku | 64.4% | estimated ± 9.3 pp, low confidence |
| 178 | Gemma 3 27B | 64.4% | estimated ± 9.3 pp, low confidence |
| 179 | Command A+ | 64.4% | estimated ± 9.3 pp, low confidence |
| 180 | LFM2.5-8B-A1B | 64.4% | estimated ± 9.3 pp, low confidence |
| 181 | Granite 4.2 3B | 64.3% | estimated ± 9.3 pp, low confidence |
| 182 | Exaone 4.0 32B | 64.0% | estimated ± 12.4 pp, low confidence |
| 183 | Sarvam 105B | 63.9% | estimated ± 12.4 pp, low confidence |
| 184 | Ling 3.0 Flash | 63.6% | estimated ± 7.2 pp, medium confidence |
| 185 | Qwen3.5-35B-A3B | 63.4% | measured |
| 186 | Qwen3 235B 2507 | 62.6% | measured |
| 187 | o3-pro | 62.5% | estimated ± 10.1 pp, low confidence |
| 188 | QwQ-32B | 62.4% | estimated ± 2.0 pp, low confidence |
| 189 | Grok 3 Mini | 61.7% | estimated ± 8.5 pp, low confidence |
| 190 | Claude 4.1 Opus | 61.7% | estimated ± 10.1 pp, low confidence |
| 191 | o1-preview | 61.7% | estimated ± 10.1 pp, low confidence |
| 192 | Claude 3 Opus | 61.7% | estimated ± 10.1 pp, low confidence |
| 193 | DeepSeek R1 Distill Qwen 32B | 61.7% | estimated ± 10.1 pp, low confidence |
| 194 | Gemini 1.5 Pro | 61.7% | estimated ± 10.1 pp, low confidence |
| 195 | Gemini 1.5 Flash | 61.7% | estimated ± 10.1 pp, low confidence |
| 196 | GPT-4 Turbo | 61.7% | estimated ± 10.1 pp, low confidence |
| 197 | Qwen2.5 Coder 32B Instruct | 61.7% | estimated ± 10.1 pp, low confidence |
| 198 | GPT-4o mini | 61.7% | estimated ± 10.1 pp, low confidence |
| 199 | DeepSeek-V2.5 | 61.7% | estimated ± 10.1 pp, low confidence |
| 200 | Gemini 1.0 Pro | 61.7% | estimated ± 10.1 pp, low confidence |
| 201 | Sarvam 30B | 61.6% | estimated ± 12.4 pp, low confidence |
| 202 | Ornith-1.5-9B | 61.3% | estimated ± 7.2 pp, medium confidence |
| 203 | Claude 4 Opus Thinking | 60.7% | estimated ± 8.5 pp, low confidence |
| 204 | LongCat-Flash-Lite-Sparse | 60.6% | estimated ± 12.4 pp, low confidence |
| 205 | Qwen3 32B (Reasoning) | 60.5% | estimated ± 2.0 pp, low confidence |
| 206 | Gemini 2.5 Pro | 59.9% | estimated ± 7.2 pp, medium confidence |
| 207 | o1-pro | 59.9% | estimated ± 8.5 pp, low confidence |
| 208 | Gemma 4 12B | 59.7% | estimated ± 8.5 pp, low confidence |
| 209 | K-EXAONE 2.0 | 59.5% | estimated ± 7.2 pp, medium confidence |
| 210 | Nemotron 3 Super 100B | 59.4% | estimated ± 7.2 pp, medium confidence |
| 211 | Claude 3.7 Sonnet Thinking | 58.9% | estimated ± 8.5 pp, low confidence |
| 212 | Gemma 4 26B A4B | 58.4% | estimated ± 7.2 pp, medium confidence |
| 213 | Qwen3 235B-A22B | 58.3% | estimated ± 2.0 pp, low confidence |
| 214 | Ternary Bonsai 2 27B | 58.3% | estimated ± 4.1 pp, high confidence |
| 215 | o4-mini | 58.1% | estimated ± 8.5 pp, low confidence |
| 216 | o3-mini | 57.6% | estimated ± 8.5 pp, low confidence |
| 217 | Kimi K2 | 57.2% | measured |
| 218 | Celeris-1 | 56.3% | estimated ± 12.4 pp, low confidence |
| 219 | GPT-OSS 120B | 56.3% | estimated ± 7.2 pp, medium confidence |
| 220 | Nemotron Ultra 253B | 56.1% | estimated ± 8.5 pp, low confidence |
| 221 | MiniMax-Text-01 | 56.1% | estimated ± 12.4 pp, low confidence |
| 222 | Grok 2 | 55.8% | estimated ± 12.4 pp, low confidence |
| 223 | o1 | 55.7% | estimated ± 8.5 pp, low confidence |
| 224 | Mellum2.1-12B-A2.5B-Thinking | 55.6% | estimated ± 4.1 pp, high confidence |
| 225 | Nemotron 3.5 Lightning 30B A3B NVFP4 | 55.6% | estimated ± 8.5 pp, low confidence |
| 226 | Grok 3 [Beta] | 55.4% | estimated ± 8.5 pp, low confidence |
| 227 | GLM-4.7-Flash | 55.1% | estimated ± 8.5 pp, low confidence |
| 228 | Claude 4 Opus | 54.8% | estimated ± 8.5 pp, low confidence |
| 229 | DeepSeek V3.1 | 54.8% | estimated ± 8.5 pp, low confidence |
| 230 | Trinity-Large-Thinking | 53.0% | estimated ± 13.9 pp, low confidence |
| 231 | GPT-OSS 20B | 52.5% | estimated ± 7.2 pp, medium confidence |
| 232 | Mellum2-12B-A2.5B-Thinking | 52.4% | estimated ± 4.1 pp, high confidence |
| 233 | Nemotron 3 Nano 30B | 52.2% | estimated ± 7.2 pp, medium confidence |
| 234 | Nemotron 3 Nano Omni 30B A3B | 51.7% | estimated ± 8.5 pp, low confidence |
| 235 | Grok 2 mini | 51.6% | estimated ± 12.4 pp, low confidence |
| 236 | Qwen3 32B | 51.3% | estimated ± 4.1 pp, high confidence |
| 237 | DeepSeek-R1 | 50.9% | estimated ± 8.5 pp, low confidence |
| 238 | ZAYA1-8B | 50.4% | estimated ± 8.5 pp, low confidence |
| 239 | Magistral Medium | 50.2% | estimated ± 8.5 pp, low confidence |
| 240 | Spark-X2.6-4B | 49.3% | estimated ± 8.5 pp, low confidence |
| 241 | LFM2.5-24B-A2B | 49.2% | estimated ± 7.2 pp, medium confidence |
| 242 | LFM3-1.7B | 49.0% | estimated ± 7.2 pp, medium confidence |
| 243 | GPT-4.5 | 48.9% | estimated ± 8.5 pp, low confidence |
| 244 | Phi-4-reasoning-plus | 48.3% | estimated ± 8.5 pp, low confidence |
| 245 | Llama 3.3 70B | 48.0% | estimated ± 12.4 pp, low confidence |
| 246 | Qwen2.5-VL-32B | 47.9% | estimated ± 12.4 pp, low confidence |
| 247 | Magistral Small | 47.7% | estimated ± 8.5 pp, low confidence |
| 248 | Claude 3.7 Sonnet | 47.5% | estimated ± 8.5 pp, low confidence |
| 249 | ZAYA1-74B-Preview | 47.1% | estimated ± 12.4 pp, low confidence |
| 250 | Granite 4.2 30B | 46.1% | estimated ± 8.5 pp, low confidence |
| 251 | GPT-4.1 | 46.0% | estimated ± 8.5 pp, low confidence |
| 252 | Phi-4-reasoning | 45.6% | estimated ± 8.5 pp, low confidence |
| 253 | Llama 3.1 70B | 45.3% | estimated ± 12.4 pp, low confidence |
| 254 | DeepSeek R1 Distill Llama 70B | 45.1% | estimated ± 8.5 pp, low confidence |
| 255 | Gemini 2.0 Flash | 45.1% | estimated ± 8.5 pp, low confidence |
| 256 | Gemini 2.5 Flash-Lite | 44.6% | estimated ± 8.5 pp, low confidence |
| 257 | GPT-4.1 mini | 44.3% | estimated ± 8.5 pp, low confidence |
| 258 | Granite 4.2 8B | 44.3% | estimated ± 8.5 pp, low confidence |
| 259 | DeepSeek-R1-0528-Qwen3-8B | 42.2% | estimated ± 8.5 pp, low confidence |
| 260 | Qwen2.5-Max | 41.6% | estimated ± 8.5 pp, low confidence |
| 261 | o1-mini | 41.6% | estimated ± 8.5 pp, low confidence |
| 262 | Claude 3.5 Sonnet | 41.3% | estimated ± 8.5 pp, low confidence |
| 263 | DeepSeek R1 Distill Qwen 14B | 41.1% | estimated ± 8.5 pp, low confidence |
| 264 | DeepSeek V3 | 41.1% | estimated ± 8.5 pp, low confidence |
| 265 | Ling 2.6 Flash | 41.0% | estimated ± 8.5 pp, low confidence |
| 266 | Gemma 4 E4B | 40.8% | estimated ± 8.5 pp, low confidence |
| 267 | MiniCPM5-2B | 40.8% | measured |
| 268 | Gemma 4 E2B | 38.6% | estimated ± 12.4 pp, low confidence |
| 269 | Mellum2-12B-A2.5B-Instruct | 36.0% | estimated ± 4.1 pp, high confidence |
| 270 | Trinity-Large-Preview | 34.8% | estimated ± 13.9 pp, low confidence |
| 271 | Soofi S 30B-A3B | 30.6% | estimated ± 12.4 pp, low confidence |
| 272 | Llama 3.1 8B | 27.9% | estimated ± 12.4 pp, low confidence |
| 273 | Phi-3.5-mini | 27.1% | estimated ± 12.4 pp, low confidence |
| 274 | LLaDA2.2-mini | 23.7% | estimated ± 13.9 pp, low confidence |
| 275 | MiniCPM5-1B | 23.1% | measured |
| 276 | Granite-4.0-1B | 16.7% | estimated ± 12.4 pp, low confidence |
| 277 | Granite-4.0-H-1B | 15.9% | estimated ± 12.4 pp, low confidence |
| 278 | LFM2.5-230M | 7.6% | estimated ± 12.4 pp, low confidence |
| 279 | LFM2.5-350M | 7.4% | estimated ± 12.4 pp, low confidence |
| 280 | LFM2.5-VL-450M | 7.0% | estimated ± 12.4 pp, low confidence |
| 281 | Granite-4.0-H-350M | 4.1% | estimated ± 12.4 pp, low confidence |
| 282 | Granite-4.0-350M | 2.8% | estimated ± 12.4 pp, low confidence |
All leaderboards
Agentic · terminal
- AA Terminal-Bench 2.1
- AA Terminal-Bench 4.0
- Terminal-Bench 4.0
- Terminal-Bench 2.0
- Terminal-Bench 2.1
- Terminal-Bench 3.0
- Terminal-Bench Hard
- Terminal-Bench-Science 0.1
- Terminal-Bench 4.1
- Terminal-Bench 2.1 (Vals AI)
Agentic · tools
- AA Agentic Index
- AA-AnalystAgent
- AA AutomationBench
- CWE-Bench-AA
- AA Cyber Index
- DeepsecBench-AA
- AA EnterpriseOps-Gym
- GDP.pdf
- AA Harvey LAB v1.0
- Harvey LAB-AA v1.1
- AA ITBench
- AA Tau3 Banking
- Agents' Last Exam
- APEX-Agents
- APEX-Agents-AA
- ApprenticeBench
- AutomationBench
- BFCL v4
- BrowseComp
- Claw-Eval
- CWE-bench v1
- CWE-bench
- CyberGym
- DeepPlanning
- DeepSearchQA
- DRACO
- ExploitGym
- Finance Agent v2
- GDPval-AA
- Gert Labs
- Hermes Bench
- Hermes Index
- HLE w/ tools
- JobBench
- LifeVal
- MazeBench
- MCP Atlas
- MCP-Tasks
- MKA
- OSWorld 2.1
- OSWorld 2.1 Offline
- OSWorld-Verified
- OSWorld 2.0
- PinchBench
- QwenClawBench
- ResearchClawBench
- SkillsBench
- τ²-bench results
- τ³-bench results
- Toolathlon
- Toolathlon-Verified
- VITA-Bench
- WideResearch
Coding
- AA Coding Index
- AA LiveCodeBench
- AA-SciCode
- Bug Hunt Bench
- CursorBench 3.1
- CursorBench 3.2
- CursorBench 4.0
- cursorBench41
- DeepSWE
- FrontierCode 1.1 Main
- FrontierCode 2.0
- FrontierCode 1.1 Extended
- FrontierSWE v2
- LiveCodeBench
- LiveCodeBench Pro
- LiveCodeBench v6
- NL2Repo
- OpenHarmony Bench v1.2
- OpenHarmony Bench v1.0
- PostTrainBench
- PostTrainBench v1
- PostTrainBench v1.1
- ProgramBench
- React Native Evals
- ScarfBench
- SciCode
- SWE-Atlas Codebase QnA
- SWE-bench Pro
- SWE-bench Verified
- SWE-Rebench
- SWE-Marathon
- Vals LiveCodeBench
- Vals SWE-bench
- Vibe Code Bench
- VulcanBench Frontier v4
- VulcanBench v3
Math
- AIME 2024
- AIME 2025
- AIME26
- FrontierMath v2 (Tier 4)
- FrontierMath v2 (Tiers 1-3)
- HMMT Feb 2025
- HMMT Feb 2026
- HMMT Nov 2025
- IMOAnswerBench
- MATH-500
- MMAnswerBench
Knowledge & reasoning
- AA-GPQA Diamond
- HealthBench Professional (length-adjusted, Artificial Analysis)
- HealthBench Professional (raw, Artificial Analysis)
- AA-HLE
- AA MMLU-Pro
- ARC-AGI-2
- ARC-AGI-3 Provider Adapter
- ARC-AGI-1
- ARC-AGI-3
- Artificial Analysis Intelligence Index
- BBH
- BioBasicsBench
- BioMysteryBench (human-difficult)
- BioMysteryBench (human-solvable)
- C-Eval
- CritPt
- GPQA
- GPQA Diamond
- HealthBench (raw)
- HealthBench Hard
- HealthBench (length-adjusted)
- HealthBench Professional
- HealthBench Professional (raw)
- HLE
- HLE w/o tools
- HLE-Verified
- JevBench 1.4
- LABBench2
- MedXpertQA (Text)
- MMLU
- MMLU-Pro
- MMLU-Pro (Arcee)
- MMLU-Redux
- MMMLU
- AA-Omniscience Accuracy
- SimpleQA
- STEMBench
- SuperGPQA
- Vals GPQA Diamond
- Vals MMLU-Pro
Instruction following
Multilingual
Vision & documents
- AA-MMMU-Pro
- AI2D_TEST
- BabyVision
- BabyVision w/ Python
- BenchCAD Vision2Code (no tools)
- BenchCAD Vision2Code (tools)
- CC-OCR
- Chartography (no tools)
- Chartography (tools)
- CharXiv
- CharXiv w/o tools
- CountBench
- DynaMath
- ERQA
- LVBench
- MathVision
- MathVision w/ Python
- MedXpertQA (MM)
- MLVU (M-Avg)
- MMMU
- MMMU-Pro
- MMMU-Pro w/ Python
- MMVU
- OCRBench V2
- OfficeQA
- OfficeQA Pro
- OmniDocBench 1.5
- ParseBench-1
- ParseBench-2
- PerceptionBench
- RealWorldQA
- RefCOCO (avg)
- ScreenSpot Pro
- SimpleVQA
- SWE-bench Multimodal
- Video-MME (w/o subtitle)
- Video-MME (with subtitle)
- VideoMMMU
- V*
- We-Math
- ZeroBench
- ZeroBench w/ Python
Long context
General
- HLE-MultiChoice (Decision Index)
- Artificial Analysis Decision Index v1.0
- Artificial Analysis Decision Index
- When2Call (Decision Index)
- WildGuardTest (Decision Index)
- BabaIsBench
- BinaryAudit
- BlenderWorld (text-only)
- Chess Puzzles
- CompileBench
- Cookie-Bench
- DAYJOB
- DAYJOB Healthcare
- Decision Index Full Score
- EBR-bench
- ExploitBench
- ExploitBench Internal Port
- FACTS Multimodal
- FailureSensorIQ
- FrontierPhysics
- Furniture Assembly
- Decision Index Arts
- Decision Index Knowledge
- Decision Index Language
- Decision Index Retrieval
- Decision Index Tools
- MaxxBench
- MirrorCode
- Mystery Game Puzzles
- NYT Connections
- OTelBench
- SEC-Bench Pro
- SkillsBench official leaderboard (BenchFlow)
- SpireClimb
- StealthMark