| # | Model | Provider | Type | Score | |
|---|---|---|---|---|---|
| 1 | Qwen3.6 Plus | Alibaba | Closed | qwen3-6-plus | |
| 2 | DeepSeek V4 Pro Base | DeepSeek | Open | deepseek-v4-pro-base | |
| 3 | Qwen3.5 397B | Alibaba | Open | qwen3-5-397b | |
| 4 | Claude Opus 4.5 | Anthropic · Closed | 92.2% | ||
| 5 | DeepSeek V4 Flash Base | DeepSeek · Open weight | 92.1% | ||
| 6 | Qwen3.6-27B | Alibaba · Open weight | 91.4% | ||
| 7 | Qwen3.6-35B-A3B | Alibaba · Open weight | 90% |