| # | Model | Provider | Type | Score | |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.5 | Anthropic | Closed | 96.6% | |
| 2 | Qwen3.7 Max | Alibaba | Closed | 95% | |
| 3 | Qwen3.5 397B | Alibaba | Open | 94.9% | |
| 4 | Qwen3.6 Plus | Alibaba | Closed | 94.5% | |
| 5 | Qwen3.6-27B | Alibaba | Open | 93.5% | |
| 6 | DeepSeek V4 Pro Base | DeepSeek | Open | 90.8% | |
| 7 | DeepSeek V4 Flash Base | DeepSeek | Open | 89.4% | |
| 8 | MiniCPM5-1B | OpenBMB | Open | 70.1% |