Qwen3.8 Max
AlibabaRank #3 of 11 · 30 PRs · 39/95 golden bugs found
42.4
F1 · +4.9 vs avg
Recall
41.0%+5.8 vs avg
Precision
43.9%-0.5 vs avg
Cost / PR
$1.210
Cost / bug found
$0.93
Confidence
95% bootstrap interval on recall (2000 resamples over the 30 PRs):29.6–53.8points. This measures sampling variance from which PRs are in the set.
Judge run-to-run variance not yet measured for this model.
Neither measures model run-to-run variance — re-running the review agent itself, not just the judge. One pass per entry.
Run configuration
HarnesskodusAccess pathapiExecution modereplayReasoningvendor-defaultJudgeclaude-haiku-4-5
Severity mix
What the model called its own findings — not recall by severity, goldens aren't severity-tagged.
Low 24Medium 49High 20Critical 4
Category mix
Only bug/performance/security are consistent across models — the rest is free text, bucketed as other.
Bug104Performance1Security1Other1