Summary
On August 7, 2026 GitHub made Copilot code review effort levels generally available, letting teams match review depth to pull-request complexity. The former Low and Medium tiers were renamed Lite and Balanced, can be chosen per review, set as organization-wide defaults, and are shown in timeline events and PR comments.
What changed
Copilot code review effort levels went GA with two tiers: Lite for straightforward changes and Balanced for larger, complex, or sensitive changes. Low/Medium were renamed to Lite/Balanced with existing configs carried forward. Reviewers can pick the level per review, admins can set organization defaults that repositories inherit, and each review now displays which effort level was used. Available on Copilot Pro, Pro+, Max, Business, and Enterprise.
Why it matters
Uniform AI review wastes budget on trivial diffs and under-scrutinizes risky ones; tunable effort levels let teams spend deeper analysis where it matters and keep routine PRs fast and cheap. Organization defaults and visible level labels add governance and auditability to AI-assisted review, which matters as automated review scales across enterprises.
Evidence excerpt
Effort levels: Lite for straightforward changes and Balanced for larger, more complex, or sensitive changes; users can choose the level per review and admins can set organization-wide defaults.