| # | Model | Provider | Type | Score | |
|---|---|---|---|---|---|
| 1 | Claude Opus 5 | Anthropic | Closed | claude-opus-5 | |
| 2 | Claude Opus 4.8 | Anthropic | Closed | claude-opus-4-8 | |
| 3 | Kimi K3 | Moonshot AI | Closed | kimi-3 | |
| 4 | Claude Fable 5 | Anthropic · Closed | 57.9% | ||
| 5 | GPT-5.5 | OpenAI · Closed | 54.1% | ||
| 6 | GPT-5.4 | OpenAI · Closed | 53.2% | ||
| 7 | MiniMax M3 | MiniMax · Open weight | 45.1% | ||
| 8 | Claude Opus 4.7 (Adaptive) | Anthropic · Closed | 43.6% |