ReviewBench: An open benchmark for AI code review

GitHub launched ReviewBench, an open benchmark for evaluating code review agents using real pull requests and production-aligned metrics.

Summary written automatically by AgentGid from the source's own text. Check the original before relying on it.