Open Code Review AACR-Bench - Recall: leaderboard
Recall on Open Code Review's AACR-Bench run: the share of the 1,505 expert-verified issues that a reviewer's comments find. Rewards exhaustive reviewers; read beside precision, since a harness can buy recall with noise.
Metric: Recall (%). Source: open-codereview.ai. Saturation forecast: Around 2029. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 28.9 | #67 |
| 2 | Qwen 3.7 Max | 23.37 | #85 |
| 3 | GLM-5.2 | 15.9 | #80 |
| 4 | GLM-5.1 | 15.7 | #109 |
| 5 | GPT-5.5 | 15.5 | #33 |
| 6 | DeepSeek V4 Pro | 12.7 | #112 |
| 7 | Claude Opus 4.8 | 12.7 | #39 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-code-review-aacr-bench-recall · How It Works · Data refreshed daily, snapshot 2026-11-05.