Reasoning
21 scenarios across 2 subcategories. 202 models scored.
Subcategories
| Subcategory | Avg Score | Best Model | Best Score |
|---|---|---|---|
| Deduction | 75.58% | GPT-6.1 Sol (Reasoning, Medium) | 100.00% |
| Attention | 82.75% | Ember-1 (Reasoning, Max) | 96.75% |
Model Leaderboard
All models ranked by their Reasoning category score.
| # | Model | Reasoning | Deduction | Attention | Overall |
|---|---|---|---|---|---|
| 1 | GPT-6.1 Sol (Reasoning, Medium) | 97.06% | 100.00% | 94.11% | 95.88% |
| 2 | GPT-5.6 Sol (Reasoning, Medium) | 96.86% | 98.99% | 94.73% | 95.15% |
| 3 | Gemini 3.8 Flash (Reasoning, Medium) | 96.72% | 97.11% | 96.34% | 95.72% |
| 4 | Grok 4.7 (Reasoning, High) | 96.71% | 99.01% | 94.41% | 92.62% |
| 5 | GPT-6 Sol (Reasoning, Medium) | 96.65% | 99.44% | 93.86% | 92.63% |
| 6 | Gemini 3.7 Flash (Reasoning, Medium) | 95.85% | 97.33% | 94.38% | 94.85% |
| 7 | Grok 4.6 (Reasoning, High) | 95.61% | 97.19% | 94.03% | 92.76% |
| 8 | Ember-1 (Reasoning, Max) | 95.50% | 94.25% | 96.75% | 95.52% |
| 9 | Claude Opus 5.5 (Reasoning) | 95.45% | 94.61% | 96.29% | 96.43% |
| 10 | Gemini 4 Argon (Reasoning, Medium) | 95.15% | 95.91% | 94.39% | 93.81% |
| 11 | GPT-5.4 (Reasoning) | 94.89% | 95.95% | 93.83% | 93.85% |
| 12 | Gemini 3.6 Flash (Reasoning) | 94.48% | 93.67% | 95.29% | 94.90% |
| 13 | Grok 4.5 (Reasoning, High) | 94.45% | 95.35% | 93.54% | 94.12% |
| 14 | DeepSeek V4 Flash 0731 (Reasoning, High) | 94.38% | 95.44% | 93.32% | 87.97% |
| 15 | GPT-6 Luna (Reasoning, High) | 94.22% | 95.74% | 92.71% | 93.95% |
| 16 | Muse Wonder (Reasoning, Medium) | 93.96% | 92.02% | 95.90% | 95.47% |
| 17 | Qwen 3.8 Max (Reasoning, XHigh) | 93.80% | 93.11% | 94.50% | 96.21% |
| 18 | Qwen 3.8 Flash (Reasoning) | 93.76% | 92.37% | 95.14% | 94.24% |
| 19 | Claude Opus 4.6 (Reasoning) | 93.19% | 90.64% | 95.74% | 95.06% |
| 20 | Z.AI GLM 5.3 (Reasoning, Max) | 93.04% | 89.81% | 96.27% | 95.20% |
| 21 | GPT-6 Sol | 92.95% | 95.61% | 90.30% | 90.05% |
| 22 | DeepSeek V4.1 Flash (Reasoning, High) | 92.95% | 92.14% | 93.76% | 94.30% |
| 23 | GPT-5.5 (Reasoning, Medium) | 92.87% | 92.17% | 93.57% | 93.72% |
| 24 | Claude Sonnet 5.5 (Reasoning) | 92.69% | 91.09% | 94.30% | 92.77% |
| 25 | Claude Opus 4.8 (Reasoning) | 91.76% | 88.50% | 95.02% | 92.33% |
| 26 | Kimi K2.6 | 91.69% | 90.61% | 92.77% | 92.57% |
| 27 | Muse Spark 1.3 (Reasoning, Medium) | 91.65% | 90.53% | 92.76% | 94.06% |
| 28 | GPT-5.6 Sol | 91.25% | 89.65% | 92.84% | 90.76% |
| 29 | GPT-5.6 Terra (Reasoning, Medium) | 91.03% | 89.85% | 92.21% | 92.49% |
| 30 | Z.AI GLM 5.1 | 90.64% | 87.24% | 94.04% | 93.74% |
| 31 | GPT-6 Luna (Reasoning, Medium) | 90.50% | 90.25% | 90.75% | 92.40% |
| 32 | Z.AI GLM 5.3 Flash (Reasoning, Max) | 90.49% | 86.93% | 94.05% | 94.28% |
| 33 | Kimi K3 (Reasoning, High) | 90.48% | 87.24% | 93.71% | 92.99% |
| 34 | Xiaomi MIMO v2.6 Flash | 90.43% | 89.54% | 91.32% | 87.31% |
| 35 | Mistral Large 4.0 (Reasoning) | 90.24% | 90.28% | 90.20% | 92.14% |
| 36 | DeepSeek V4 Pro 0813 (Reasoning, High) | 90.05% | 87.16% | 92.93% | 90.21% |
| 37 | Qwen 3.8 27B (Reasoning, XHigh) | 89.96% | 89.83% | 90.10% | 92.91% |
| 38 | Hy4 Preview (Reasoning, High) | 89.85% | 87.80% | 91.90% | 94.80% |
| 39 | Claude Opus 4.6 | 89.75% | 86.15% | 93.35% | 92.31% |
| 40 | Claude Sonnet 4.6 (Reasoning) | 89.61% | 85.00% | 94.22% | 93.64% |
| 41 | Grok 4.5 (Reasoning, Low) | 89.59% | 86.44% | 92.75% | 90.94% |
| 42 | Kimi K3 (Reasoning, Low) | 89.46% | 86.14% | 92.77% | 92.07% |
| 43 | GPT-5 | 89.34% | 85.13% | 93.55% | 91.48% |
| 44 | Gemini 2.5 Pro | 89.21% | 86.38% | 92.04% | 88.44% |
| 45 | Muse Spark 1.1 (Reasoning, Medium) | 89.03% | 85.67% | 92.40% | 93.81% |
| 46 | Claude Opus 5 (Reasoning, Low) | 88.95% | 86.32% | 91.59% | 92.06% |
| 47 | GPT-5.5 (Reasoning, Low) | 88.87% | 87.08% | 90.66% | 92.51% |
| 48 | Claude Opus 4.8 (Reasoning, Low) | 88.83% | 85.19% | 92.47% | 91.89% |
| 49 | Muse Spark 1.2 (Reasoning, Medium) | 88.79% | 84.03% | 93.54% | 94.71% |
| 50 | Claude Opus 4.7 (Reasoning) | 88.52% | 85.06% | 91.99% | 92.53% |
| 51 | Claude Opus 4.7 | 88.29% | 85.68% | 90.91% | 89.90% |
| 52 | Gemini 3.5 Flash (Reasoning) | 88.22% | 85.49% | 90.96% | 93.35% |
| 53 | Gemini 3.1 Pro (Preview) | 88.20% | 86.52% | 89.89% | 94.08% |
| 54 | Xiaomi MIMO v2.6 Pro | 88.14% | 84.33% | 91.96% | 91.50% |
| 55 | Kimi K2.5 | 87.99% | 86.07% | 89.91% | 90.86% |
| 56 | Claude Sonnet 5 (Reasoning, Low) | 87.78% | 83.40% | 92.17% | 90.16% |
| 57 | Claude Sonnet 5 (Reasoning) | 87.77% | 83.41% | 92.12% | 90.40% |
| 58 | GPT-5.5 | 87.68% | 85.39% | 89.98% | 89.37% |
| 59 | GPT-5.4 (Reasoning, Low) | 86.99% | 83.06% | 90.91% | 90.91% |
| 60 | Z.AI GLM 5 Turbo | 86.87% | 83.04% | 90.71% | 93.29% |
| 61 | Aion 3.0 | 86.56% | 84.03% | 89.10% | 88.78% |
| 62 | MiniMax M3 | 86.27% | 83.04% | 89.51% | 90.45% |
| 63 | Gemini 3 Flash (Preview, Reasoning) | 86.16% | 83.35% | 88.97% | 89.93% |
| 64 | Aion 3.5 Mini (Reasoning, High) | 86.15% | 82.66% | 89.65% | 92.36% |
| 65 | GPT-5.4 Mini (Reasoning) | 86.10% | 82.62% | 89.58% | 89.82% |
| 66 | Thinking Machines Inkling (Reasoning) | 86.03% | 81.28% | 90.78% | 90.78% |
| 67 | Z.AI GLM 5.2 (Reasoning, High) | 85.81% | 80.97% | 90.66% | 93.41% |
| 68 | Qwen 3.6 Max Preview | 85.79% | 79.68% | 91.90% | 93.72% |
| 69 | Claude Opus 5 | 85.64% | 82.80% | 88.48% | 89.97% |
| 70 | GPT-5.2 | 85.61% | 81.54% | 89.68% | 89.45% |
| 71 | DeepSeek V4 Flash (Reasoning) | 85.45% | 83.08% | 87.83% | 88.06% |
| 72 | Claude Opus 5 (Reasoning) | 85.32% | 82.56% | 88.08% | 91.15% |
| 73 | Muse Spark 1.1 (Reasoning, Minimal) | 85.18% | 80.90% | 89.46% | 91.43% |
| 74 | Claude Haiku 5.5 (Adaptive) | 85.05% | 79.31% | 90.78% | 89.37% |
| 75 | Aion 3.5 (Reasoning, High) | 85.04% | 81.80% | 88.28% | 91.90% |
| 76 | GPT-5.6 Terra | 84.98% | 80.74% | 89.22% | 88.23% |
| 77 | GPT-5.6 Luna (Reasoning, Medium) | 84.74% | 82.29% | 87.19% | 90.61% |
| 78 | Claude Opus 4.5 | 84.25% | 77.87% | 90.63% | 89.60% |
| 79 | Z.AI GLM 5 | 84.09% | 79.21% | 88.97% | 89.60% |
| 80 | DeepSeek V4 Pro (Reasoning) | 83.52% | 76.94% | 90.10% | 89.28% |
| 81 | Qwen 3.7 Max | 83.47% | 83.59% | 83.35% | 94.55% |
| 82 | Gemini 3.6 Flash (Reasoning, Minimal) | 83.40% | 81.93% | 84.87% | 86.68% |
| 83 | GPT-5 Mini | 83.31% | 78.44% | 88.18% | 91.31% |
| 84 | GPT-5.1 | 82.57% | 79.33% | 85.81% | 90.73% |
| 85 | o4 Mini High | 82.51% | 77.69% | 87.32% | 88.78% |
| 86 | Z.AI GLM 4.7 | 82.43% | 79.52% | 85.33% | 87.67% |
| 87 | Qwen 3.5 397B A17B | 81.97% | 75.79% | 88.15% | 91.09% |
| 88 | Qwen 3.5 Plus (2026-02-15) | 81.81% | 75.68% | 87.94% | 86.17% |
| 89 | Xiaomi MIMO v2.5 | 81.80% | 74.69% | 88.92% | 83.95% |
| 90 | ByteDance Seed 2.0 Lite | 81.47% | 75.38% | 87.56% | 84.27% |
| 91 | Claude Haiku 5.5 (Reasoning, Low) | 81.20% | 73.68% | 88.71% | 84.75% |
| 92 | Z.AI GLM 4.6 | 81.01% | 77.79% | 84.23% | 87.64% |
| 93 | Gemini 3 Flash (Preview) | 80.99% | 75.02% | 86.97% | 85.47% |
| 94 | GPT-5.4 | 80.92% | 75.68% | 86.16% | 84.31% |
| 95 | Qwen 3.5 Plus (2026-04-20) | 80.60% | 73.17% | 88.02% | 89.79% |
| 96 | o4 Mini | 80.39% | 76.48% | 84.29% | 86.56% |
| 97 | Gemma 4 31B (Reasoning) | 80.14% | 74.02% | 86.25% | 89.64% |
| 98 | Qwen 3.6 27B | 79.84% | 72.95% | 86.74% | 88.33% |
| 99 | Thinking Machines Inkling Small (Reasoning, High) | 79.84% | 73.22% | 86.46% | 87.39% |
| 100 | ByteDance Seed 2.0 Mini | 79.82% | 72.65% | 86.98% | 85.69% |
| 101 | ByteDance Seed 1.6 | 79.79% | 74.00% | 85.57% | 89.59% |
| 102 | Xiaomi MIMO v2.5 Pro | 79.73% | 73.13% | 86.34% | 86.05% |
| 103 | Claude Haiku 5.5 (Reasoning, Medium) | 79.69% | 71.09% | 88.29% | 84.17% |
| 104 | Claude Sonnet 4.6 | 79.62% | 72.09% | 87.14% | 90.66% |
| 105 | Qwen 3.6 Flash | 79.51% | 71.49% | 87.53% | 89.31% |
| 106 | Qwen 3.5 27B | 79.44% | 76.50% | 82.38% | 90.05% |
| 107 | Qwen 3.5 Flash | 79.34% | 71.85% | 86.84% | 85.66% |
| 108 | Qwen 3.5 122B | 79.24% | 71.05% | 87.43% | 90.32% |
| 109 | Gemini 2.5 Flash | 79.09% | 74.16% | 84.03% | 80.61% |
| 110 | Qwen 3.7 Flash (Reasoning) | 78.94% | 75.40% | 82.48% | 88.56% |
| 111 | Gemini 3.5 Flash (Reasoning, Minimal) | 78.93% | 72.73% | 85.12% | 85.88% |
| 112 | Z.AI GLM 5.3 Flash (Reasoning, Low) | 78.82% | 72.55% | 85.10% | 88.76% |
| 113 | Aion 2.0 | 78.64% | 74.98% | 82.30% | 86.66% |
| 114 | Claude Sonnet 4.5 | 78.61% | 70.37% | 86.85% | 87.54% |
| 115 | Gemma 4 31B | 78.18% | 69.61% | 86.75% | 85.23% |
| 116 | GPT-5.4 Nano (Reasoning) | 77.96% | 76.88% | 79.04% | 80.02% |
| 117 | Qwen 3.5 35B | 77.89% | 68.75% | 87.03% | 87.01% |
| 118 | Nemotron 3 Super | 77.87% | 75.92% | 79.82% | 81.69% |
| 119 | Z.AI GLM 4.5 | 77.00% | 72.94% | 81.07% | 84.95% |
| 120 | GPT-OSS 120B | 77.00% | 71.54% | 82.46% | 84.81% |
| 121 | GPT-6 Luna | 76.94% | 72.46% | 81.43% | 83.73% |
| 122 | Claude Opus 4 | 76.72% | 67.28% | 86.16% | 87.22% |
| 123 | Muse Glimmer 30B (Reasoning, Medium) | 76.61% | 69.08% | 84.13% | 88.81% |
| 124 | Qwen 3.6 35B | 76.37% | 68.80% | 83.95% | 87.66% |
| 125 | Inception Mercury 2 | 76.23% | 71.53% | 80.92% | 81.99% |
| 126 | Mistral Large 4.0 | 76.15% | 70.72% | 81.58% | 84.72% |
| 127 | Claude Sonnet 4 | 75.86% | 67.21% | 84.52% | 87.64% |
| 128 | Aion 3.0 Mini | 75.86% | 80.61% | 71.10% | 83.69% |
| 129 | Mistral Large 2 | 75.76% | 69.57% | 81.95% | 81.50% |
| 130 | Gemini 3.1 Flash Lite (Preview) | 75.75% | 66.64% | 84.86% | 85.41% |
| 131 | Grok 4.3 (Reasoning) | 75.64% | 68.02% | 83.26% | 90.99% |
| 132 | DeepSeek V3.2 | 75.48% | 69.80% | 81.17% | 82.22% |
| 133 | GPT-4o, Aug. 6th (temp=0) | 75.40% | 72.24% | 78.55% | 82.18% |
| 134 | Claude Sonnet 5 | 75.35% | 66.73% | 83.97% | 87.34% |
| 135 | Gemini 3.1 Flash Lite (Reasoning) | 75.32% | 66.95% | 83.69% | 85.91% |
| 136 | Mistral Large 3 | 75.26% | 69.47% | 81.06% | 84.29% |
| 137 | Mistral Medium 3.5 (Reasoning) | 74.65% | 66.87% | 82.43% | 85.32% |
| 138 | Gemini 3.1 Flash Lite | 74.62% | 66.65% | 82.59% | 85.09% |
| 139 | GPT-5.4 Mini (Reasoning, Low) | 74.50% | 66.36% | 82.64% | 83.57% |
| 140 | GPT-4.1 | 74.40% | 68.81% | 79.99% | 86.82% |
| 141 | MiniMax M2.5 | 74.33% | 71.72% | 76.93% | 86.71% |
| 142 | Grok 4.20 (Reasoning) | 74.28% | 60.08% | 88.49% | 90.87% |
| 143 | Mistral Medium 3.5 | 74.19% | 67.37% | 81.00% | 84.52% |
| 144 | Gemini 3.5 Flash Lite (Reasoning) | 74.16% | 71.41% | 76.90% | 84.34% |
| 145 | Gemma 4 26B (Reasoning) | 74.05% | 68.86% | 79.24% | 89.02% |
| 146 | Gemma 4 26B | 73.83% | 66.56% | 81.09% | 84.89% |
| 147 | Gemini 3.5 Flash Lite (Reasoning, Minimal) | 73.22% | 71.77% | 74.67% | 84.45% |
| 148 | GPT-5.6 Luna | 72.88% | 66.67% | 79.09% | 85.06% |
| 149 | MiniMax M2.7 | 72.76% | 67.57% | 77.95% | 86.23% |
| 150 | Gemini 2.5 Flash (Reasoning) | 72.63% | 72.79% | 72.48% | 84.14% |
| 151 | Gemini 2.5 Flash Lite (Reasoning) | 72.51% | 71.64% | 73.38% | 83.10% |
| 152 | Grok 4.20 | 72.46% | 66.13% | 78.80% | 81.21% |
| 153 | DeepSeek V3.1 | 72.08% | 67.31% | 76.85% | 82.35% |
| 154 | Mistral Medium 3.1 | 71.84% | 69.60% | 74.07% | 76.08% |
| 155 | DeepSeek V4 Pro | 71.72% | 61.62% | 81.82% | 82.05% |
| 156 | Qwen 3.5 9B | 70.90% | 65.74% | 76.05% | 84.05% |
| 157 | DeepSeek-V2 Chat | 70.75% | 62.62% | 78.87% | 84.09% |
| 158 | GPT-4o, Aug. 6th (temp=1) | 70.58% | 66.52% | 74.64% | 81.28% |
| 159 | GPT-5 Nano | 70.50% | 66.00% | 75.00% | 80.16% |
| 160 | Grok 4.3 | 70.17% | 60.21% | 80.13% | 78.00% |
| 161 | GPT-5.4 Mini | 70.07% | 64.78% | 75.36% | 80.45% |
| 162 | Z.AI GLM 4.7 Flash | 69.68% | 67.90% | 71.45% | 82.21% |
| 163 | Writer: Palmyra X5 | 69.64% | 63.44% | 75.85% | 78.11% |
| 164 | Laguna S 2.1 | 69.62% | 62.23% | 77.01% | 79.99% |
| 165 | DeepSeek V3 (2025-03-24) | 69.32% | 66.68% | 71.96% | 79.93% |
| 166 | DeepSeek V3 (2024-12-26) | 69.12% | 63.28% | 74.95% | 82.62% |
| 167 | DeepSeek V4 Flash | 68.68% | 53.69% | 83.66% | 82.02% |
| 168 | Z.AI GLM 4.5 Air | 68.46% | 60.93% | 75.99% | 80.74% |
| 169 | Qwen 3 235B A22B Instruct 2507 | 68.43% | 62.64% | 74.23% | 78.07% |
| 170 | Claude Haiku 4.5 | 67.77% | 58.43% | 77.12% | 83.36% |
| 171 | Ministral 3 14B | 67.43% | 68.13% | 66.73% | 70.45% |
| 172 | GPT-4.1 Mini | 67.35% | 60.26% | 74.43% | 81.40% |
| 173 | Laguna XS 2.1 | 67.05% | 62.87% | 71.22% | 75.42% |
| 174 | Mistral Small 4 (Reasoning) | 66.73% | 62.74% | 70.72% | 79.48% |
| 175 | Qwen 3 32B | 66.35% | 64.44% | 68.27% | 79.37% |
| 176 | Gemini 2.5 Flash Lite | 65.19% | 63.84% | 66.54% | 79.91% |
| 177 | Nemotron 3 Nano | 65.05% | 68.20% | 61.91% | 74.50% |
| 178 | Thinking Machines Inkling | 64.85% | 54.93% | 74.76% | 73.96% |
| 179 | Hermes 3 405B | 64.59% | 59.50% | 69.68% | 80.80% |
| 180 | GPT-5.4 Nano (Reasoning, Low) | 64.43% | 65.86% | 63.00% | 77.46% |
| 181 | Mistral Small 3.2 24B | 63.87% | 57.06% | 70.67% | 77.36% |
| 182 | ByteDance Seed 1.6 Flash | 63.50% | 67.00% | 60.00% | 70.92% |
| 183 | Ministral 8B | 63.20% | 58.16% | 68.24% | 63.77% |
| 184 | GPT-4o Mini (temp=1) | 62.26% | 62.17% | 62.34% | 77.82% |
| 185 | GPT-4o Mini (temp=0) | 62.17% | 60.87% | 63.47% | 76.86% |
| 186 | Qwen 2.5 72B | 61.71% | 57.16% | 66.25% | 73.17% |
| 187 | Mistral Small 4 | 61.67% | 61.55% | 61.79% | 75.23% |
| 188 | Gemma 3 27B | 61.51% | 58.39% | 64.63% | 75.70% |
| 189 | Llama 3.1 70B | 61.45% | 55.78% | 67.13% | 77.41% |
| 190 | WizardLM 2 8x22b | 60.41% | 56.40% | 64.41% | 71.45% |
| 191 | Ministral 3 8B | 59.53% | 50.65% | 68.42% | 69.98% |
| 192 | GPT-5.4 Nano | 58.53% | 59.95% | 57.12% | 72.16% |
| 193 | Cydonia 24B V4.1 | 56.95% | 58.13% | 55.78% | 72.68% |
| 194 | Gemma 3 12B | 56.28% | 72.50% | 40.06% | 76.07% |
| 195 | Hermes 3 70B | 56.20% | 57.86% | 54.53% | 69.74% |
| 196 | Arcee AI: Trinity Mini | 55.45% | 54.53% | 56.36% | 67.68% |
| 197 | GPT-4.1 Nano | 54.55% | 61.83% | 47.26% | 69.90% |
| 198 | Ministral 3 3B | 54.06% | 50.62% | 57.49% | 65.02% |
| 199 | Ministral 3B | 53.21% | 49.47% | 56.95% | 59.25% |
| 200 | Gemma 3 4B | 52.86% | 60.65% | 45.06% | 66.33% |
| 201 | Cohere Command R+ (Aug. 2024) | 51.60% | 51.97% | 51.23% | 67.04% |
| 202 | Mistral NeMO | 44.87% | 52.31% | 37.43% | 63.80% |