SWE-PRBench: leaderboard
Pull-request review benchmark with a public paper-baseline leaderboard for model review quality and false-positive behavior.
Metric: Overall (sbar) (imported). Source: huggingface.co. Source: benchmarklist.com. Saturation forecast: Around April 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Haiku 4.5 | 15.3 | #280 |
| 2 | Claude Sonnet 4.6 | 15.2 | #90 |
| 3 | DeepSeek V3 | 15 | #321 |
| 4 | Mistral Large 3 | 14.7 | #401 |
| 5 | GPT-4o | 11.3 | #343 |
| 6 | GPT-4o Mini | 10.8 | #599 |
| 7 | Mistral Small | 10.6 | #740 |
| 8 | Llama 3.3 70B | 7.9 | #579 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=swe-prbench · How It Works · Data refreshed daily, snapshot 2026-10-20.