benchmark · first seen 18 Oct 2026
code review benchmark
Evaluation suites measuring how well models or agents review code changes and find defects.
Mentions per week
PapersCodeDiscussion
Earliest sources
- From Software History to Executable Regression Tasks: A System for Evaluating Software AgentsPapers · 18 Oct 2026 · Ruikai Shi, Zhengran Zeng, Rui Xie, Shikun Zhang, Wei Ye
Latest sources
- From Software History to Executable Regression Tasks: A System for Evaluating Software AgentsPapers · 18 Oct 2026 · Ruikai Shi, Zhengran Zeng, Rui Xie, Shikun Zhang, Wei Ye