cursor Bench31
6 models evaluated
|
| 1 | Claude Opus 5.5 | Anthropic | Proprietary | 57.8% | |
| 2 | Claude Fable 5.1 | Anthropic | Proprietary | 51.8% | |
| 3 | Grok 4.7 | xAI | Proprietary | 46.3% | |
| 4 | GPT-5.6 Sol | OpenAI · Closed | 41.7% | | |
| 5 | Gemini 3.8 Flash | Google · Closed | 39.6% | | |
| 6 | Claude Sonnet 5 | Anthropic · Closed | 34.1% | | |