| Rank | Model | Provider | Score |
|---|---|---|---|
| 1 | Claude Opus 5 | Anthropic | 65.7% |
| 2 | MiMo V2.6 Pro | Xiaomi | 62.0% |
| 3 | MiMo V2.6 Flash | Xiaomi | 61.2% |
| 4 | Claude Fable 5 | Anthropic | 57.4% |
| 5 | Qwen3.8 Flash Next | Alibaba | 55.7% |
| 6 | Kimi K3 | Moonshot AI | 54.3% |
| 7 | Qwen3.8 Max | Alibaba | 53.4% |
| 8 | DeepSeek V4.1 Flash | DeepSeek | 45.8% |
| 9 | GPT 5.6 Sol | OpenAI | 45.4% |
| 10 | Qwen3.8 27B | Alibaba | 33.4% |
| 11 | MiMo V2.5 Pro | Xiaomi | 25.0% |
| 12 | Qwen3.6 27B | Alibaba | 21.8% |
Figures are taken from public model cards, technical reports and vendor announcements, and are approximate where sources disagree. Results from different benchmark versions or tool-use settings never share a column: Terminal Bench 2.1, 3.0 and DeepSWE v1.1 are distinct benchmarks, and HLE with tools is separate from HLE without tools. Only models with a published result are listed; the score column is shaded relative to the best and worst result on this benchmark.