| 1 | Claude Fable 5 | Anthropic | Proprietary | 74.15% | |
| 2 | Claude Opus 5 | Anthropic | Proprietary | 73.90% | |
| 3 | Kimi K3 | Moonshot AI | Proprietary | 73.42% | |
| 4 | GPT-5.6 Sol | OpenAI · Closed | max reasoning | 72.64% | |
| 5 | Claude Opus 4.8 | Anthropic · Closed | 70.89% | | |
| 6 | Muse Spark 1.2 | Meta · Closed | 69.80% | | |
| 7 | GPT-5.6 Luna | OpenAI · Closed | max reasoning | 69.06% | |
| 8 | Claude Sonnet 5 | Anthropic · Closed | 68.83% | | |
| 9 | GPT-5.5 | OpenAI · Closed | xhigh reasoning | 68.07% | |
| 10 | Claude Opus 4.7 | Anthropic · Closed | 67.36% | | |
| 11 | Muse Spark 1.1 | Meta · Closed | xhigh reasoning | 66.74% | |
| 12 | Qwen3.8 Max | Alibaba · Open weight | 65.39% | | |
| 13 | Gemini 3.6 Flash | Google · Closed | high reasoning | 65.08% | |
| 14 | GPT-5.6 Terra | OpenAI · Closed | xhigh reasoning | 65.07% | |
| 15 | Grok 4.5 | xAI · Closed | high reasoning | 63.42% | |
| 16 | Gemini 3.5 Flash | Google · Closed | high reasoning | 62.93% | |
| 17 | Claude Sonnet 4.6 | Anthropic · Closed | 60.57% | | |
| 18 | MiniMax M3 | MiniMax · Open weight | 59.97% | | |
| 19 | Kimi K2.6 | Moonshot AI · Open weight | 56.43% | | |
| 20 | Gemini 3.1 Pro Preview | Google | high reasoning | 56.07% | |
| 21 | GPT-5.4 mini | OpenAI · Closed | xhigh reasoning | 54.22% | |
| 22 | Qwen3.7 Plus | Alibaba · Closed | 53.89% | | |
| 23 | MiMo-V2.5 | Xiaomi · Closed | 52.77% | | |
| 24 | Gemini 3 Flash Preview | Google | high reasoning | 52.19% | |
| 25 | Qwen3.6 Plus | Alibaba · Closed | 51.52% | | |
| 26 | Inkling-Small | Thinking Machines Lab · Open weight | 0.99 reasoning | 50.05% | |
| 27 | Inkling | Thinking Machines Lab · Open weight | 0.99 reasoning | 49.40% | |
| 28 | GPT-5.4 nano | OpenAI · Closed | high reasoning | 47.64% | |
| 29 | Grok 4.3 | xAI · Closed | high reasoning | 43.29% | |
| 30 | Claude Haiku 4.5 Thinking | Anthropic · Closed | 42.88% | | |
| 31 | Gemini 3.1 Flash Lite Preview | Google | high reasoning | 41.35% | |
| 32 | Grok 4.20 0309 Reasoning | xAI | 39.06% | | |
| 33 | Mistral Medium 3.5 | Mistral AI | high reasoning | 34.77% | |