Israeli Startup Baz Tops New Independent AI Code Review Benchmark, Outperforming OpenAI and Google

Summary

Israeli startup Baz has ranked first in precision on the newly launched Code Review Bench, an independent benchmark designed specifically to evaluate AI-powered code review systems. The benchmark positions itself as the first methodologically transparent, independent comparison focused on code review quality, distinguishing itself from existing benchmarks like SWE-Bench that have faced criticism for models being trained to optimize directly against them. Code Review Bench combines controlled eva