用Claude审Codex代码，效率提升显著；反向则效果下降。
Cross-Model LLM Code Review: Should you use Claude to review Codex or vice versa?
- 测试不同模型配对的代码审查效果，模拟真实开发流程。
- Claude审Codex使通过率从71.6%升至89.7%，显著提升。
- 反向配对（Codex审Claude）导致通过率下降，不推荐使用。
开发者日益采用两个编码代理协作：一个生成草稿，另一个进行审查。但这种搭配是否值得时间成本，以及顺序是否重要尚不明确。我们在116个近期的hard和medium难度的LCB任务上，对Claude和Codex进行了六种条件的受控实验，以模拟软件开发者的实际工作流：两种单模型基线、两种跨模型配对顺序、两种同模型配对顺序。审查者可见问题描述和编写者的草稿，但无法执行测试，这更贴近真实代码审查环节。结果显示，由Claude审查Codex的草稿，可将通过率从71.6%提升至89.7%（p_BH = .001）；Codex自审可提升至84.5%（p_BH = .022）。而反向操作效果不佳：Codex审查Claude的草稿，通过率从91.4%降至82.8%（p_BH = .046），Claude自审则保持91.4%不变。结果表明，有效配对具有不对称性：应使用Claude审查Codex，而非相反。
原文摘要 · Abstract (English)
Developers increasingly use two coding agents together: one writes a draft, and the other reviews it. However, it is not clear whether the pairing is worth its cost and time, or whether the order of the pairing matters. We run a controlled experiment on 116 recent hard and medium lcb tasks with Claude and Codex across six conditions to approximate a software practitioner's workflow: both solo baselines, both cross-model orderings, and both same-model orderings. The reviewer sees the problem and the writer's draft but cannot execute tests, which approximates a code review step. Claude review raises Codex drafts from 71.6% to 89.7% ($p_{BH}=.001$); Codex self review raises them to 84.5% ($p_{BH}=.022$). The reverse direction does not pay off: Codex reviewing Claude drafts drops the pass rate from 91.4% to 82.8% ($p_{BH}=.046$), and Claude self review leaves the 91.4% baseline unchanged. Our evaluation indicates that the useful pairing is asymmetric: use Claude to review Codex, not the other way around.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。