GitHub announced on October 2, 2026 that GitHub Copilot code reviews can be requested through REST and GraphQL interfaces, with an effort level selected for the request. The change is generally available on Pro, Pro+, Max, Business and Enterprise plans. It lets a team start a review from a script or an internal system, instead of relying solely on a manual action in the interface.
The review request can come from within the workflow
According to the changelog, selecting an effort level in an API request is optional, not a field that must be set every time. For a team building automation, this means the request can be integrated at the point where the decision that a change is ready for review is already made.
For example, an internal process could be designed to request a review after the project's initial tests finish. This is a possible use case, not a sequence activated automatically by the update. The developer still needs to define the conditions, permissions and handling of a failed request.
The same announcement also mentions an earlier change: since September 28, Default uses Balanced. An earlier explicit selection of Lite was retained. These are two dates for two different details; the new development on October 2 is API access, and the default change should not be presented as having taken effect only then.
When designing automation, it is also advisable to determine what will happen if a review is delayed or fails. A team can display a clear status to the developer and leave the decision on how to proceed with the responsible person. Silence from an automated process should not be interpreted as confirmation that no problems were found in the code.
Review depth and automatic activation are not the same switch
In GitHub's documentation, Lite is described as a standard review, while Balanced is intended for deeper analysis of complex logic, security-sensitive code and changes spanning services. The Max option is marked Coming soon and is not yet available. The Max plan name does not make the Max review level available.
GitHub says Balanced consumes more AI credits and may use slightly more GitHub Actions minutes. The default setting should therefore be examined as part of the work budget, without assuming that additional depth comes with no change in consumption.
The documentation separates selecting the effort level from enabling automatic reviews. Turning off automatic review does not erase the effort level selected for manual reviews. In addition, automatic requests may originate from user settings or repository and organization rules; a personal setting does not necessarily override an organizational rule.
For a team wanting to try the API, a good starting point is a test repository with a small change and an expected result. Verify that the review was actually requested, check which level it ran at and go through the comments before expanding usage. A successful API call confirms that the system received a request, but does not prove that the code is correct. Responsibility for deciding whether the change is ready to merge remains part of the development process.