Observed facts: Anthropic says Claude Code Review is generally available to Claude Team and Enterprise organizations, and that most reviews now take about 6–10 minutes rather than roughly 20 minutes during preview. Mechanism: Repository-level Quick, Standard, and Deep settings let teams trade review depth against turnaround according to the codebase, while automatic checks on new pull requests add review capacity without requiring developers to initiate each run.
Why now: As coding agents increase the pace of code production, a review queue can absorb or erase that speedup; Anthropic explicitly frames the persistent review bottleneck as the problem. The product’s new availability and tuning controls make that operational response usable beyond a preview cohort.
Implication: Teams can test whether automated review catches useful issues before human review and whether faster feedback reduces queue time. Teams should measure false positives and missed defects rather than equating review speed with quality. Evidence boundary: The announcement reports Anthropic’s own performance and quality claims and does not provide independent accuracy measurements, comparative evaluations, or evidence of reduced merge time.
Watch condition: Look for independently measured defect detection, false-positive rates, and review-queue effects across repositories using the different depth settings.