Epoch AI - Proofbench — leaderboard
Metric: Score. Source: epoch.ai. 43 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol (Max) | 77 |
| 2 | Claude Fable 5 (Max) | 77 |
| 3 | GPT-5.6 Terra (xHigh) | 71 |
| 4 | Claude Opus 4.8 (Max) | 69 |
| 5 | Claude Sonnet 5 (Max) | 66 |
| 6 | GPT-5.4 (xHigh) | 56 |
| 7 | GPT-5.6 Luna (Max) | 54 |
| 8 | Claude Opus 4.7 (Max) | 54 |
| 9 | GPT-5.5 (xHigh) | 50 |
| 10 | Claude Opus 4.6 (Max) | 50 |
| 11 | Claude Sonnet 4.6 (Max) | 45 |
| 12 | Muse Spark 1.1 | 39 |
| 13 | Claude Opus 4.5 (20251101) (Unknown) | 36 |
| 14 | GLM-5.2 (Max) | 35 |
| 15 | Grok 4.5 (High) | 30 |
Interactive version: aibenchmarks.dev/benchmark?slug=epoch-ai-proofbench · How the rankings work · Data refreshed daily, snapshot 2026-07-16.