| 1 | Claude Fable 5.1 | Anthropic | Proprietary | 90.5% | |
| 2 | Claude Fable 5 | Anthropic | Proprietary | 89.8% | |
| 3 | Gemini 3.8 Flash | Google | Proprietary | high reasoning | |
| 4 | Claude Opus 5 | Anthropic · Closed | 89.0% | | |
| 5 | Gemini 3.7 Flash | Google · Closed | high reasoning | 88.7% | |
| 6 | Gemini 3.1 Pro Preview | Google | high reasoning | 88.5% | |
| 7 | Grok 4.6 | xAI · Closed | high reasoning | 88.2% | |
| 8 | Gemini 3.6 Flash | Google · Closed | high reasoning | 88.1% | |
| 9 | GPT-5.2-Codex | OpenAI · Closed | 88.0% | | |
| 10 | Qwen3.8 Max | Alibaba · Open weight | 87.9% | | |
| 11 | Claude Opus 4.8 | Anthropic · Closed | 87.8% | | |
| 12 | Gemini 3.5 Flash | Google · Closed | high reasoning | 87.6% | |
| 13 | DeepSeek V4 Pro 0813 | DeepSeek · Closed | max reasoning | 87.5% | |
| 15 | Grok 4.5 | xAI · Closed | high reasoning | 87.4% | |
| 16 | GPT-5.3 Codex | OpenAI · Closed | xhigh reasoning | 87.3% | |
| 17 | DeepSeek V4 Flash 0731 | DeepSeek · Closed | high reasoning | 87.3% | |
| 18 | Kimi K3 | Moonshot AI · Closed | max reasoning | 87.2% | |
| 19 | Qwen3.7 Max | Alibaba · Closed | 87.1% | | |
| 20 | Kimi K2.6 | Moonshot AI · Open weight | 86.8% | | |
| 21 | GPT-5 mini | OpenAI · Closed | high reasoning | 86.6% | |
| 22 | GPT-5.1 | OpenAI · Closed | high reasoning | 86.5% | |
| 23 | Gemini 3 Pro Preview | Google | 86.4% | | |
| 24 | Nemotron 3 Ultra 550b A55b | Nvidia | 86.0% | | |
| 25 | Qwen3.6 Plus | Alibaba · Closed | 86.0% | | |
| 26 | Inkling-Small | Thinking Machines Lab · Open weight | 0.99 reasoning | 85.9% | |
| 27 | GPT-5.6 Terra | OpenAI · Closed | xhigh reasoning | 85.9% | |
| 28 | GPT-5 | OpenAI | high reasoning | 85.9% | |
| 29 | Muse Spark 1.1 | Meta · Closed | xhigh reasoning | 85.9% | |
| 30 | Gemini 3 Flash Preview | Google | high reasoning | 85.6% | |
| 31 | GPT-5.1-Codex | OpenAI · Closed | 85.5% | | |
| 32 | Inkling | Thinking Machines Lab · Open weight | 0.99 reasoning | 85.5% | |
| 33 | GPT-5.2 | OpenAI · Closed | xhigh reasoning | 85.4% | |
| 34 | Qwen3.5 Plus Thinking | Alibaba | 85.3% | | |
| 35 | GPT-5.5 | OpenAI · Closed | xhigh reasoning | 85.3% | |
| 36 | Claude Opus 4.7 | Anthropic · Closed | 85.1% | | |
| 37 | GPT-5 Codex | OpenAI | high reasoning | 84.7% | |
| 38 | Claude Opus 4.6 (Adaptive) | Anthropic · Closed | 84.7% | | |
| 39 | Grok 4.3 | xAI · Closed | high reasoning | 84.5% | |
| 40 | Grok 4.20 0309 Reasoning | xAI | 84.3% | | |
| 41 | GPT-5.4 | OpenAI · Closed | xhigh reasoning | 84.1% | |
| 42 | GPT-5.4 nano | OpenAI · Closed | high reasoning | 84.0% | |
| 43 | Qwen3.8-27B | Alibaba · Open weight | xhigh reasoning | 84.0% | |
| 44 | Ling 3.0 Flash 2607 | Ant | 84.0% | | |
| 45 | o3 | OpenAI · Closed | high reasoning | 83.9% | |
| 46 | Kimi K2.5 Thinking | Moonshot AI | 83.9% | | |
| 47 | Claude Opus 4.5 Thinking | Anthropic · Closed | 83.7% | | |
| 48 | GPT-5.1-Codex-Max | OpenAI · Closed | high reasoning | 83.6% | |
| 49 | Qwen3.5 Flash | Alibaba · Closed | 83.3% | | |
| 50 | Grok 4 0709 | xAI | 83.2% | | |
| 51 | GPT-OSS 120B | OpenAI · Open weight | 83.2% | | |
| 52 | GPT-5.6 Sol | OpenAI · Closed | max reasoning | 82.6% | |
| 53 | Claude Sonnet 5 | Anthropic · Closed | 82.4% | | |
| 54 | GLM-4.7 | Z.AI · Open weight | 82.2% | | |
| 55 | O4 Mini | OpenAI | high reasoning | 82.2% | |
| 56 | MiniMax M3 | MiniMax · Open weight | 82.2% | | |
| 57 | Claude Sonnet 4.6 | Anthropic · Closed | 82.1% | | |
| 58 | Kimi K2.7 Code | Moonshot AI · Open weight | 82.0% | | |
| 59 | GLM 5 Thinking | Zhipu AI | 81.9% | | |
| 60 | MiniMax M2.1 | MiniMax | 81.8% | | |
| 61 | MiMo-V2.5 | Xiaomi · Closed | 81.5% | | |
| 62 | GPT-5.4 mini | OpenAI · Closed | xhigh reasoning | 81.5% | |
| 63 | GLM-5.1 | Z.AI · Open weight | 81.4% | | |
| 64 | MiMo-V2.5-Pro | Xiaomi · Closed | 81.4% | | |
| 65 | GLM-4.6 | Z.AI · Open weight | 81.0% | | |
| 66 | DeepSeek V3p2 Thinking | Fireworks AI | 80.7% | | |
| 67 | Grok 4.1 Fast (Reasoning) | xAI · Closed | 80.6% | | |
| 68 | GLM-5.3 | Z.AI · Open weight | max reasoning | 80.5% | |
| 69 | GLM-5.3-Flash | Z.AI · Open weight | max reasoning | 80.5% | |
| 70 | GPT-OSS 20B | OpenAI · Open weight | 80.4% | | |
| 71 | Gemini 3.1 Flash Lite Preview | Google | 80.1% | | |
| 72 | MiniMax M2.7 | MiniMax · Open weight | 79.9% | | |
| 73 | MiniMax M2.5 | MiniMax · Closed | 79.2% | | |
| 74 | Gemini 2.5 Pro Preview 03 25 | Google | 79.2% | | |
| 75 | Gemini 3.5 Flash-Lite | Google · Closed | high reasoning | 79.0% | |
| 76 | Grok 4 Fast (Reasoning) | xAI · Closed | 79.0% | | |
| 77 | Qwen3 Max | Alibaba · Closed | 78.2% | | |
| 78 | Grok 3 Mini Fast High Reasoning | xAI | high reasoning | 76.2% | |
| 79 | Gemini 2.5 Flash Preview 09 2025 Thinking | Google | 76.2% | | |
| 80 | Gemini 2.5 Flash Preview 09 2025 | Google | 75.1% | | |
| 81 | Claude Opus 4.5 | Anthropic · Closed | 75.0% | | |
| 82 | Magistral Medium 2509 | Mistral AI | 74.9% | | |
| 83 | Claude Sonnet 4.5 Thinking | Anthropic · Closed | 73.0% | | |
| 84 | Magistral Small 2509 | Mistral AI | 72.1% | | |
| 85 | o3-mini | OpenAI · Closed | high reasoning | 71.5% | |
| 86 | Gemini 2.5 Flash Lite Preview 09 2025 Thinking | Google | 71.4% | | |
| 87 | Qwen3 235b A22b | Fireworks AI | 70.6% | | |
| 88 | Moonshotai Kimi K2 Instruct | Together AI | 70.4% | | |
| 89 | DeepSeek-R1 | DeepSeek · Open weight | 70.2% | | |
| 90 | GPT-5 nano | OpenAI · Closed | high reasoning | 70.2% | |
| 91 | Claude Opus 4 20250514 Thinking | Anthropic | 70.2% | | |
| 92 | DeepSeek V3p2 | Fireworks AI | 69.9% | | |
| 93 | GLM-5.2 | Z.AI · Open weight | 69.5% | | |
| 94 | Laguna M.1 | Poolside · Closed | 68.1% | | |
| 95 | Laguna XS.2 | Poolside · Open weight | 67.8% | | |
| 96 | Gemini 2.5 Flash Lite Preview 09 2025 | Google | 67.7% | | |
| 97 | GLM-4.5 | Z.AI · Closed | 67.4% | | |
| 98 | Qwen3 Max Preview | Alibaba | 66.9% | | |
| 99 | Claude Opus 4.1 20250805 Thinking | Anthropic | 66.5% | | |
| 100 | Grok 3 Mini Fast Low Reasoning | xAI | low reasoning | 66.3% | |
| 101 | DeepSeek V3 0324 | Fireworks AI | 65.5% | | |
| 102 | Claude Opus 4.1 | Anthropic | 64.6% | | |
| 103 | Kimi K2 Thinking | Moonshot AI | 63.1% | | |
| 104 | Claude Opus 4 | Anthropic | 62.6% | | |
| 105 | Claude Sonnet 4 20250514 Thinking | Anthropic | 62.4% | | |
| 106 | Grok Code Fast 1 | xAI · Closed | 62.0% | | |
| 107 | Claude 3.7 Sonnet 20250219 Thinking | Anthropic | 60.4% | | |
| 108 | Claude Sonnet 4 | Anthropic | 59.7% | | |
| 109 | Langston Nim Nvidia Llama 3.3 Nemotron Super 49b V1 42e84561 Thinking | Together AI | 58.4% | | |
| 110 | GPT-4.1 mini | OpenAI · Closed | high reasoning | 58.2% | |
| 111 | Gemini 2.5 Flash Preview 04 17 | Google | 56.9% | | |
| 112 | Claude 3.7 Sonnet | Anthropic | 56.7% | | |
| 113 | Mistral Large 2512 | Mistral AI | 55.3% | | |
| 114 | GPT-4.1 | OpenAI · Closed | high reasoning | 54.7% | |
| 115 | Grok 3 | xAI | 52.9% | | |
| 116 | Devstral 2512 | Mistral AI | 51.8% | | |
| 117 | o1 | OpenAI · Closed | high reasoning | 50.3% | |
| 118 | Claude 3.5 Sonnet | Anthropic · Closed | 49.6% | | |
| 119 | Llama4 Maverick Instruct Basic | Fireworks AI | 47.3% | | |
| 120 | Gemini 2.5 Flash Preview 04 17 Thinking | Google | 46.9% | | |
| 121 | Grok 4 Fast Non Reasoning | xAI | 46.1% | | |
| 122 | Mistral Medium 2505 | Mistral AI | 44.8% | | |
| 123 | Gemini 2.0 Flash 001 | Google | 43.6% | | |
| 124 | GPT-4o | OpenAI · Closed | high reasoning | 43.4% | |
| 125 | Labs Devstral Small 2512 | Mistral AI | 43.2% | | |
| 126 | GPT-4.1 nano | OpenAI · Closed | high reasoning | 42.7% | |
| 127 | Grok 4.1 Fast Non Reasoning | xAI | 42.6% | | |
| 128 | Claude 3.5 Haiku | Anthropic | 41.9% | | |
| 129 | Gemini 1.5 Pro 002 | Google | 41.7% | | |
| 130 | Claude Haiku 4.5 Thinking | Anthropic · Closed | 41.2% | | |
| 131 | Grok 2 1212 | xAI | 38.7% | | |
| 132 | Meta Llama Llama 4 Scout 17B 16E Instruct | Together AI | 38.5% | | |
| 133 | Mistral Large 2411 | Mistral AI | 37.1% | | |
| 134 | Gemini 1.5 Flash 002 | Google | 36.9% | | |
| 135 | Meta Llama Llama 3.3 70B Instruct Turbo | Together AI | 36.3% | | |
| 136 | Langston Nim Nvidia Llama 3.3 Nemotron Super 49b V1 42e84561 | Together AI | 36.3% | | |
| 137 | Command A 03 2025 | Cohere | 35.1% | | |
| 138 | Mistral Small 2503 | Mistral AI | 31.8% | | |
| 139 | GPT-4o mini | OpenAI · Closed | high reasoning | 26.4% | |
| 140 | Jamba Large 1.6 | AI21 Labs | 22.3% | | |
| 141 | Command R Plus | Cohere | 18.2% | | |
| 142 | Mistral Small 2402 | Mistral AI | 15.8% | | |
| 143 | Jamba Mini 1.6 | AI21 Labs | 9.9% | | |