Voices
Microsoft Launches Critique: GPT + Claude Multi-Model Deep Research in M365 Copilot
Microsoft announced Critique on March 30, a new multi-model research system in M365 Copilot where OpenAI's GPT drafts the initial research response and Anthropic's Claude reviews it for accuracy, completeness, and citation quality before the user sees it. Alongside Critique, a 'Council' mode shows multiple model responses side-by-side with a cover letter explaining where they agree and diverge. Microsoft claims Critique achieves +7.0 points over Perplexity Deep Research (Claude Opus 4.6), a 13.88% improvement.
↳ Follow the thread