Factory Code Review Benchmark: leaderboard
Factory benchmark for code review quality, scoring model comments against expected findings with mean F1 across realistic pull request review tasks.
Metric: Mean F1 (%). Source: factory.ai. Saturation forecast: Around 2029. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 60.5 | #110 |
| 2 | Claude Opus 4.6 | 59.8 | #63 |
| 3 | Claude Sonnet 4.6 | 57.4 | #90 |
| 4 | Claude Opus 4.7 | 55.9 | #49 |
| 5 | GLM-5.1 | 55.8 | #97 |
| 6 | GPT-5.3 Codex | 55.7 | #69 |
| 7 | Gemini 3.1 Pro (Preview) | 52.1 | #58 |
| 8 | Kimi K2.5 | 51.9 | #144 |
| 9 | GPT-5.4 Mini | 51.5 | #212 |
| 10 | Gemini 3 Flash | 49.5 | #99 |
| 11 | GPT-5.5 | 47.9 | #31 |
| 12 | GPT-5.4 | 47.5 | #79 |
| 13 | MiniMax-M2.7 | 45.6 | #258 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=factory-code-review-benchmark · How It Works · Data refreshed daily, snapshot 2026-10-20.