How is Copilot Code Review Max different from the existing Copilot Code Review?

The existing Copilot Code Review offers two effort levels, 'Lite' and 'Balanced'. Copilot Code Review Max is realized through a newly added effort level called 'Max', which runs multiple independent review agents on multiple AI models.

How does Copilot Code Review Max combine the output of its review agents?

An aggregator agent coordinates the outputs of the multiple review agents, resolving discrepancies and reconfirming accuracy before the results are reported as a review.

Is Copilot Code Review Max generally available?

It is currently offered as a public preview.

Large Language ModelsAI Coding AssistantsPublickey1h agoNov 4, 2026, 01:00 JST

GitHub unveils Copilot Code Review Max with top ReviewBench F1 score

Send on LINE

3 Key Points

What happened
GitHub announced Copilot Code Review Max at GitHub Universe 2026, adding a new 'Max' effort level alongside 'Lite' and 'Balanced', currently in public preview.
Why it matters
It runs multiple independent review agents on multiple AI models under an aggregator agent that resolves disagreements before reporting, aiming to catch issues a single model misses.

WHO IT HITSDevelopment teams using GitHub Copilot for pull-request reviews get a heavier, multi-model review option, while teams on 'Lite' or 'Balanced' keep the existing lighter levels. The added scrutiny is aimed at code review as AI-agent-generated code and pull requests surge.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Get these free every morning

Context & Analysis

GitHub has offered Copilot Code Review for some time, detecting bugs and returning review results to developers. The new announcement comes as code generation and pull requests driven by AI agents have increased sharply, which GitHub frames as raising the importance of code review for maintaining software quality.

The mechanism behind the 'Max' level is a departure from a single-model review. Multiple independent review agents, each running on a different AI model, are combined so that one model's strengths cover another's blind spots; an aggregator agent then reconciles disagreements and rechecks accuracy before reporting. GitHub says the result recorded the highest F1 score on ReviewBench for detecting high-severity issues in code, though the body does not give the score itself or the comparison set.

The 'Max' level sits above the existing 'Lite' and 'Balanced' effort levels, giving teams a heavier option rather than replacing the lighter ones. It is available now as a public preview, so the practical test for development teams will be how the aggregated multi-model output holds up on their own repositories.

FAQ

How is Copilot Code Review Max different from the existing Copilot Code Review?: The existing Copilot Code Review offers two effort levels, 'Lite' and 'Balanced'. Copilot Code Review Max is realized through a newly added effort level called 'Max', which runs multiple independent review agents on multiple AI models.
How does Copilot Code Review Max combine the output of its review agents?: An aggregator agent coordinates the outputs of the multiple review agents, resolving discrepancies and reconfirming accuracy before the results are reported as a review.
Is Copilot Code Review Max generally available?: It is currently offered as a public preview.

PublickeyRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.