SelfBench - vercel/next.js: leaderboard
Metric: Task success (%) on 38 tasks rebuilt from the repository's merged pull requests (hidden tests). Source: selfbench.dev. Saturation forecast: Around January 2027. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6.1 Sol (High) | 84.21 |
| 2 | GPT-6.1 Sol (Low) | 78.95 |
| 3 | Claude Opus 5.5 (Low) | 68.42 |
| 4 | GPT-6 Sol (High) | 65.79 |
| 5 | GPT-6 Sol (Low) | 60.53 |
| 6 | Kimi K3 (High) | 50 |
| 7 | GPT-6 Luna (High) | 39.47 |
| 8 | GPT-6 Luna (Low) | 21.05 |
Interactive version: theaggregate.ai/benchmark?slug=selfbench-vercel-next-js · How It Works · Data refreshed daily, snapshot 2026-10-03.