| # | Model | Provider | Type | Score | |
|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 | Anthropic | Proprietary | 67.7% | |
| 2 | Claude Opus 5 | Anthropic | Proprietary | 66.9% | |
| 3 | Claude Opus 4.8 | Anthropic | Proprietary | 66.2% | |
| 4 | Hy4 preview | Tencent · Open weight | 66.2% | ||
| 5 | Kimi K3 | Moonshot AI · Closed | 63.3% | ||
| 6 | GLM-5.3-Flash | Z.AI · Open weight | 62.4% | ||
| 7 | Step 5 Preview | StepFun · Closed | 60.3% | ||
| 8 | Claude Fable 5 | Anthropic · Closed | 57.9% | ||
| 9 | GPT-5.5 | OpenAI · Closed | 54.1% | ||
| 10 | GPT-5.4 | OpenAI · Closed | 53.2% | ||
| 11 | MiniMax M3 | MiniMax · Open weight | 45.1% | ||
| 12 | Claude Opus 4.7 (Adaptive) | Anthropic · Closed | 43.6% |