Exploitbench benchmark scores from Epoch AI's benchmark data collection.
Sourced from Epoch AI's systematic benchmark data collection covering frontier models.
Auto-generated definition — editorial description not yet written.
Higher is better.
Auto-generated definition — editorial description not yet written.
| # | Model | Score |
|---|---|---|
| 1 | Claude Mythos Preview | 9.6 |
| 2 | Claude Opus 4.7 | 3.6 |
| 3 | Gemini 3.1 Pro Preview | 3.2 |
| 4 | GLM-5.1 (Reasoning) | 2.6 |
| 5 | Kimi K2.6 | 2.4 |
| 6 | Anthropic: Claude Haiku 4.5 | 2.1 |
| 7 | MiniMax-M2.7 | 2.1 |