Stubbi
Rankings
Informal comparison

IronBee raw-model code review

Compare raw language models reviewing the same real pull requests.

Checked Oct 21, 2026IronBee · Serkan Ozal

Published leaderboard

Explore scores and evaluation settings on the publisher’s leaderboard.

Open leaderboard

Coverage

21 models, 58 configurations, three repeats; a 50-PR main set and a separate 10-PR frontier subset.

Methodology

Rerun Martian Code Review Bench with a common prompt and a calibrated three-model judging panel. Report precision, recall, F2, cost and latency.

People & organizations

2

Sources

Study codeUnderlying benchmarkHacker News discussion