Anthropic Claude Code Review Is Now Generally Available
On 15 October 2026 Anthropic moved Code Review, the automated pull-request reviewer in Claude Code, out of research preview: it is now generally available to Claude Team and Enterprise organisations, Anthropic says most reviews finish about 3x faster (roughly 20 minutes down to 6 to 10 minutes), and admins pick Quick, Standard (default) or Deep per repository
Tool Overview
Details, capabilities and usage guidance
Features
- ✓ Reviews run about 3x faster: Anthropic says most reviews now finish in 6 to 10 minutes versus about 20 minutes during the research preview. The problem it targets is that coding agents keep increasing the pace at which teams write code, so human review becomes the bottleneck
- ✓ Per-repository review depth: Anthropic says admins choose Quick, Standard (the default) or Deep for each repository, so routine codebases get a quick pass and the most critical ones get a deep review, and every level delivers high-quality reviews
- ✓ Findings state the problem, the trigger and the fix: Anthropic says the GA release gives clear, concise findings that state the problem, what triggers it and the fix in a few sentences, replacing vague warnings that engineers had to decode before acting
- ✓ Enable it in Claude Code settings and install the GitHub App: Anthropic says admins enable Code Review in Claude Code settings, install the GitHub App and choose repositories. Once enabled, reviews run automatically on new pull requests rather than being requested by hand
- ✓ Enterprise credit offer: Anthropic says eligible Claude Enterprise organisations can claim at least $10,000 in Code Review credits by 15 January 2027, expiring 30 days after claiming (third-party coverage puts the top tier at $50,000 and notes that EU data-residency configurations are excluded)
How to Use
- Check that your organisation is on a Claude Team or Enterprise plan, since Anthropic says Code Review is generally available for those two plans
- Have an admin enable Code Review in Claude Code settings and install the official GitHub App, which is what makes reviews run against your repositories
- Assign a depth per repository: Quick for everyday codebases, Standard as the default and Deep for the most critical ones, so not everything runs at the same intensity
- If you are an eligible Enterprise organisation, claim the review credits before 15 January 2027, keep in mind they expire 30 days after claiming, and confirm your organisation meets the conditions Anthropic lists
FAQ
What is Code Review?
Anthropic says Code Review is the multi-agent pull-request review system in Claude Code that runs automatically on new pull requests to catch bugs before they reach production; it moved from research preview to general availability on 15 October 2026.
How much faster is it than the preview?
Anthropic says most reviews now finish about 3x faster, down from roughly 20 minutes during the preview to 6 to 10 minutes. That is Anthropic's own figure, and what you see still depends on repository size and the depth you choose.
How do I choose between the three depths?
Anthropic says admins pick Quick, Standard (the default) or Deep per repository: Quick for routine codebases and Deep for the most critical ones, with every level delivering high-quality reviews. A safe starting point is Standard, then tune it repository by repository.
How do I enable it?
Anthropic says admins enable Code Review in Claude Code settings, install the GitHub App and select the repositories to cover; after that, reviews run automatically on new pull requests.
Is there a credit offer?
Anthropic says eligible Claude Enterprise organisations can claim at least $10,000 in Code Review credits by 15 January 2027, expiring 30 days after they are claimed, with terms applying; third-party coverage puts the ceiling at $50,000 and notes some configurations are excluded.
Does it replace human code review?
Anthropic does not claim that. It frames Code Review as faster and easier to tune per repository, and says that during the preview it was catching real issues before a human reviewer looked at the PR, so the emphasis is on getting ahead of human review rather than replacing it.