Microsoft enabled GPT to work in tandem with Claude, outperforming all competitors.

According to Mars Finance, Microsoft announced two new features for Copilot Researcher on Monday—Critique and Council—combining OpenAI's GPT and Anthropic's Claude for the same research task. Critique employs a sequential collaborative model: GPT is responsible for planning the research, retrieving data, and generating the initial draft, while Claude then acts as a rigorous reviewer, verifying factual accuracy and citation quality; Council allows the two models to generate reports independently in parallel, and a third judging model compares differences and summarizes disagreements. In the DRACO benchmark test, which covers 100 complex research tasks across 10 fields including medicine, law, and technology, Copilot with Critique scored 57.4 points, leading the second-place finisher by nearly 14%, and far exceeding the 42.7 points achieved by Claude Opus 4.6 running alone.