SelfBench - vercel/eve: leaderboard
Metric: Task success (%) on 38 tasks rebuilt from the repository's merged pull requests (hidden tests). Source: selfbench.dev. Saturation forecast: Around August 2027. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6.1 Sol (Low) | 73.68 |
| 2 | Claude Opus 5.5 (High) | 73.68 |
| 3 | GPT-6.1 Sol (High) | 71.05 |
| 4 | Claude Opus 5.5 (Low) | 71.05 |
| 5 | Kimi K3 (High) | 65.79 |
| 6 | GPT-6 Luna (High) | 50 |
| 7 | GPT-6 Luna (Low) | 34.21 |
Interactive version: theaggregate.ai/benchmark?slug=selfbench-vercel-eve · How It Works · Data refreshed daily, snapshot 2026-10-03.