Anthropic Makes Claude Code Review Generally Available

On October 15, 2026, Anthropic made Code Review generally available. This is the automated, multi-agent pull-request review system for Claude Code. It moved out of research preview. The system is now available to Claude Team and Enterprise organizations.

What the system does

Code Review runs on every new pull request. It posts findings as comments on the PR. It does not approve or block the PR itself. The review is done by multiple AI agents working together holistically. The system reviews code in a fresh context. Anthropic notes this helps because Claude will not be biased toward code it just wrote. Most reviews now take 6 to 10 minutes, according to Anthropic’s reporting. That is a reported figure, not a hard guarantee. Actual time will vary by repository size and review depth.

Admin controls and setup

Admins enable Code Review centrally in Claude Code settings. They connect it through the GitHub App. Then they pick which repositories get reviewed. There is no per-developer setup. Once enabled, reviews run automatically on every new PR in the selected repositories. Admins can also set review depth per repository. Anthropic’s developer account on X (@ClaudeDevs) says admins can pick from Quick, Standard, or Deep review levels for each repo. This gives teams control over how thorough the review is versus how fast it runs. The system requires a Claude Team or Enterprise plan. It does not work with the free or Pro tiers.

Promotional credits

The general availability launch comes with a promotional offer. Eligible Enterprise organizations get free review credits. The amount is not fully clear. One outlet reports at least $10,000. Another reports up to $50,000. The exact figure is uncertain. There is one notable exclusion. Organizations on GitHub Enterprise Cloud with EU data residency are reportedly excluded from the offer. This is likely due to data handling or compliance considerations, but Anthropic has not publicly detailed the reason.

What this means for teams

The main benefit is removing manual review overhead. A team that gets a high volume of PRs can have Claude handle the first pass. The system catches style issues, potential bugs, and logic problems. It posts those as comments. Human reviewers then focus on the higher-level design and architectural decisions. The fresh-context approach is worth noting. If Claude wrote the code, it might have a blind spot. By reviewing in a new context, the system is more likely to spot flaws in its own previous work. That is a deliberate design choice. The lack of approval or blocking capability is also intentional. The system is a reviewer, not a gatekeeper. It provides information. It does not enforce policy. Teams that want automated blocking will need to build that themselves, likely through CI rules that act on the comments.

Coverage and reception

The launch was covered on October 15 by several outlets. Unite.AI, Superpower Daily, and The Next Web all ran stories. Anthropic also made its own developer announcements. The coverage is consistent with the facts above, with the credit amount being the main point of divergence.

Practical considerations

Before enabling this, teams should think about a few things. First, the review depth setting matters. Quick reviews will be faster but less thorough. Deep reviews will take longer but catch more. There is no one-size-fits-all setting. Teams should start with Standard and adjust based on their needs. Second, the system posts comments. That means PR threads will get more noise. Teams should have a process for triaging those comments. Otherwise, developers might ignore them, which defeats the purpose. Third, the EU data residency exclusion on the promo offer is a reminder that data handling is a real concern. Teams in the EU should check whether the standard service (not the promo) has any restrictions. The promo exclusion does not necessarily mean the service itself is unavailable, but it is a signal to verify.

Bottom line

Claude Code Review is now a production tool, not a research experiment. It is available to paying Team and Enterprise customers. It offers central control, per-repo depth settings, and automated PR comments. The promo credits are a nice incentive for eligible Enterprise organizations, though the exact amount is unclear. The system does not replace human review. It augments it. For teams with high PR volume, it can save time and catch issues early. For teams with low volume, the setup cost might not be worth it. That is a decision each team will make based on their own workflow. The 6-to-10-minute review time is a practical figure. It means a PR can get feedback within a coffee break. That is fast enough to keep momentum in a development cycle. It is not instant, but it is close enough for most workflows. Anthropic has been iterating on Claude Code for a while. This GA release is a clear step toward making AI-assisted code review a standard part of the development process. The tool is now mature enough for production use. The question is how well it integrates with existing team habits holistically. That will determine its actual value.