CUDABeaver - Pass@1: leaderboard
Metric: Share of the 213 broken-start CUDA debugging tasks (%) whose first repair attempt compiles, passes the task tests and runs at least 0.7 times as fast as the reference (performance gate p=0.7), default protocol: category-aware error feedback (level L3), iterative sampling at temperature 0.7; higher is better. Source: arxiv.org. Saturation forecast: Around 2028. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.6 Plus | 15.5 |
| 2 | Gemma 4 31B (IT) | 14.7 |
| 3 | GPT-5.4 | 14.5 |
| 4 | Kimi K2.6 | 13.6 |
| 5 | Qwen 3.6 27B | 10.3 |
| 6 | GLM-4.7 | 7.5 |
Interactive version: theaggregate.ai/benchmark?slug=cudabeaver-pass-1 · How It Works · Data refreshed daily, snapshot 2026-10-07.