SelfBench - vercel/vercel: leaderboard
Metric: Task success (%) on 47 tasks rebuilt from the repository's merged pull requests (hidden tests). Source: selfbench.dev. Saturation forecast: Around July 2027. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 5.5 (High) | 80.85 |
| 2 | Kimi K3 (High) | 78.72 |
| 3 | GPT-6.1 Sol (Low) | 76.6 |
| 4 | GPT-6.1 Sol (High) | 72.34 |
| 5 | GPT-6 Luna (High) | 68.09 |
Interactive version: theaggregate.ai/benchmark?slug=selfbench-vercel-vercel · How It Works · Data refreshed daily, snapshot 2026-10-03.