GitHub's Copilot Now Matches Review Depth to Pull Request Risk

GitHub just flipped its Copilot code review depth controls to general availability, giving teams a way to match the thoroughness of an automated review to the risk of the pull request in front of it. The change introduces two named tiers, Lite and Balanced, that replace the Low and Medium labels that lived in public preview.

The core idea: not every pull request needs the same scrutiny. Documentation updates and small fixes may only need focused feedback, while complex logic, security-sensitive code, and cross-service changes benefit from deeper analysis. Instead of running the same model against a typo fix and a payment-processing refactor, you now pick the right tool for the job.

Lite vs Balanced, in practice

The two tiers differ mainly in the model behind them and the depth of context they pull in:

• Lite: Standard review that provides fast, targeted feedback on common issues such as bugs, security vulnerabilities, and style inconsistencies (default).
• Balanced: Deeper analysis of complex logic, security-sensitive code, and cross-service changes, using a higher-reasoning model.
• The tradeoff is spelled out in the docs: Balanced reviews use more AI credits and more GitHub Actions minutes than Lite.

That cost delta matters because Copilot code review uses GitHub Actions to run agentic capabilities, meaning Balanced runs burn both model credits and CI minutes on every pull request they touch.

How to turn it on

There are three levels of control, and they cascade in the way you would expect from GitHub org settings:

1. Organization-level defaults: org admins set a default under Organization settings -> Copilot -> Copilot code review, and repositories inherit it unless they configure their own.
2. Repository admins can override the org default for a specific repo.
3. You can select the review effort level in the pull request, under the Reviewers section where Copilot appears as a reviewer.

If you were already using the preview, nothing breaks. The Low and Medium effort levels introduced during public preview are now named Lite and Balanced. If you previously configured an effort level, your configuration automatically carries forward under the new name.

Auditability and availability

One small but useful addition: Copilot code review now labels which effort level ran in both timeline events and the pull request overview comment. You can see at a glance whether a review ran at Balanced or Lite, making it easy to track review depth across your repositories. That solves an obvious governance question, which is whether a security-sensitive PR actually got the deeper pass or slipped through on Lite.

On pricing, Lite and Balanced are available with Copilot Pro, Pro+, Max, Business, and Enterprise plans. The documentation also notes that Copilot code review is available for organization members without a Copilot license when enabled by an enterprise administrator.

When to reach for which

Default Lite for the fast lane: dependency bumps, doc edits, small refactors, config tweaks. Escalate to Balanced when the change touches authentication, data handling, or spans multiple services, where a higher-reasoning model has a better shot at catching the kind of subtle logic bug that a pattern-matching review would miss. The honest limit still applies here, since Copilot is not guaranteed to catch everything, and GitHub says to validate its feedback alongside human review. Effort levels tune depth and cost; they do not replace judgment about what a change touches.