GitHub announced on August 27, 2026, that Copilot code review will cover more automatically requested pull requests, including bot-authored changes and pull requests opened by Copilot cloud agent. The announcement also introduces three resolution reasons for Copilot review comments: Addressed, Won’t fix, and Incorrect.
For engineering teams, the change can extend automated review to changes that do not come directly from a developer and remove the former 300-file or 20,000-line limitation. For administrators, however, the expanded workflow depends on automatic-review configuration, policy settings, GitHub Actions runner capacity, and AI-credit billing. GitHub’s documentation does not specify a replacement maximum size for pull requests, so teams should not interpret the removed threshold as a guarantee of predictable processing for arbitrarily large changes.
What changed in Copilot code review
According to GitHub’s announcement, Copilot code review can review bot-authored pull requests when a review is requested automatically. This includes pull requests opened by Copilot cloud agent. When automatic review is configured, pull requests from the cloud agent can receive a full agentic review.
GitHub also says the previous limit of 300 files or 20,000 lines of code for a pull request no longer applies. The supplied documentation does not define a new maximum, performance target, latency expectation, or processing limit. Pull-request size still matters operationally: GitHub documents that AI-credit consumption generally increases with pull-request size and repository custom instructions.
The other announced change concerns how developers close Copilot review conversations. A Copilot code review comment can be resolved using a dropdown next to the Resolve conversation button. The available reasons are:
- Addressed — the finding was handled.
- Won’t fix — the team chose not to make the suggested change.
- Incorrect — the finding was not valid.
GitHub says these selections provide feedback to its product team and can help improve the product. The supplied sources confirm the reasons for Copilot code review comments, but do not establish whether the same taxonomy applies to every possible Copilot review comment type.
Automatic review is the key configuration point
Bot-authored pull requests are not automatically covered simply because a repository uses Copilot. Automatic review must be configured at the relevant repository, organization, or enterprise scope.
For a repository, administrators can use Settings → Rules → Rulesets, create an active branch ruleset, select the target branches, and enable Automatically request Copilot code review. Organization owners can configure rulesets for repositories selected through inclusion or exclusion patterns. Enterprise owners can configure enterprise-level branch rulesets that target organizations and repositories.
Automatic review rules can be set to review a pull request once, review new pushes, and review draft pull requests. Reviewing every new push or draft can increase coverage, but GitHub notes that these options may also create more developer noise. Organizations should decide whether the additional feedback is useful for their workflow before applying the rule broadly.
The automatic-review configuration is documented in GitHub’s setup guide. Review effort defaults can also be configured at the organization and repository levels, with repository settings overriding organization defaults.
License and policy requirements
Copilot code review is available on paid Copilot plans. A separate no-license workflow is documented for organizations using Copilot Business or Copilot Enterprise. This distinction matters for teams that want people without individual Copilot licenses to request reviews or have their pull requests covered by automatic review.
For no-license use on GitHub.com, an enterprise administrator or organization owner must first enable the AI credits paid usage policy. The organization must then enable Allow members without a Copilot license to use Copilot code review in GitHub.com. The no-license policy is disabled by default. If it is set at the enterprise level, it becomes visible at the organization level but cannot be edited there.
The policy applies only to repositories under an organization where it has been explicitly enabled. Once both policies are active, users without a Copilot license can request reviews on pull requests in the organization’s repositories. Automatic review applies to pull requests in repositories where automatic code review is enabled, regardless of whether the pull-request author has a Copilot license.
GitHub’s Copilot code review documentation describes these requirements and the available review surfaces. No-license code review is documented for GitHub.com, not for IDEs. The supplied sources do not establish every repository eligibility condition for the newly expanded bot-authored and very-large-pull-request coverage.
Costs, credits, and runner requirements
Copilot code review consumes AI credits according to the model used and the number of tokens processed. GitHub estimates $0.05 to $1 of AI credits for Lite reviews and $0.25 to $5 for Balanced reviews. These are estimates, not fixed prices for every review.
Copilot Business includes 1,900 AI credits per user per month, while Copilot Enterprise includes 3,900. Usage beyond the shared enterprise pool is charged at $0.01 per AI credit. Teams enabling no-license reviews should therefore treat the policy as a billing decision as well as a permissions decision. The relevant plan allowances and usage details are listed in GitHub’s Copilot plans documentation.
Agentic capabilities use GitHub Actions runners. Standard GitHub-hosted runners are used by default. Larger hosted runners can improve performance at a higher per-minute rate, while self-hosted runners do not consume GitHub Actions minutes. If GitHub-hosted runners are disabled, agentic capabilities are unavailable and reviews fall back to a more limited mode. A review may still be generated if Actions workflows fail, but without the additional agentic features.
What developers and administrators should do
- Review eligibility and billing. Confirm whether the organization is on Copilot Business or Enterprise before planning no-license review access. Evaluate AI-credit usage and possible additional charges before enabling the policy.
- Enable policies in the required order. For no-license use, enable AI credits paid usage first, then enable the no-license Copilot code-review policy. The supplied documentation does not provide the exact settings navigation for the second policy.
- Configure automatic review at the right scope. Use repository, organization, or enterprise rulesets depending on which repositories and branches should receive reviews. Include new pushes or draft pull requests only when the resulting review volume is appropriate.
- Plan runner capacity. Check whether GitHub-hosted runners are available for agentic reviews and determine whether larger or self-hosted runners fit the team’s performance and cost requirements.
- Use resolution reasons consistently. Select Addressed, Won’t fix, or Incorrect when resolving Copilot review comments so the outcome is recorded and can contribute product feedback.
- Keep human review in the process. GitHub warns that Copilot can make mistakes and is not guaranteed to identify every problem. Findings, particularly in security-sensitive or complex changes, should be validated by developers.
Important limitations and availability details
Copilot code review supports code written in any language, but GitHub excludes dependency management files such as package.json and Gemfile.lock, along with log and SVG files. Lite and Balanced review efforts provide different analysis depth and have different estimated credit consumption.
The announcement removes the earlier pull-request size threshold, but neither the announcement nor the supplied current documentation states a replacement maximum. Teams handling very large changes should monitor processing behavior and costs rather than assume that all large pull requests will receive identical performance.
The announcement is dated August 27, 2026. The supplied sources do not specify rollout stages or regional availability for the expanded capabilities. GitHub’s documentation lists GitHub.com, GitHub CLI, GitHub Mobile, VS Code, Visual Studio, Xcode, JetBrains IDEs, and Azure DevOps public preview as supported Copilot code-review surfaces; no-license review is documented for GitHub.com only.
