SelfBench - mupt-ai/self-bench: leaderboard
Metric: Task success (%) on 38 tasks rebuilt from the repository's merged pull requests (hidden tests). Source: selfbench.dev. Saturation forecast: Around June 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6.1 Sol (High) | 76.32 | #9 (GPT-6.1 Sol) |
| 2 | Claude Opus 5.5 (High) | 71.05 | #3 (Claude Opus 5.5) |
| 3 | GPT-6.1 Sol (Low) | 65.79 | #9 (GPT-6.1 Sol) |
| 4 | Kimi K3 (High) | 63.16 | #25 (Kimi K3) |
| 5 | GLM-5.3 (High) | 52.63 | #36 (GLM-5.3) |
| 6 | GPT-6 Luna (High) | 47.37 | #49 (GPT-6 Luna) |
No result here: #5 GPT-6 Astra, #8 Claude Fable 5.1, #10 Claude Sonnet 5.5.
Interactive version: theaggregate.ai/benchmark?slug=selfbench-mupt-ai-self-bench · How It Works · Data refreshed daily, snapshot 2026-10-11.