Microsoft Copilot Now Uses GPT and Claude Simultaneously

Microsoft is taking a bold step to make its AI assistant even more capable. The company has announced that Copilot is getting a significant upgrade in its research capabilities by combining the strengths of two industry leaders: OpenAI’s ChatGPT and Anthropic’s Claude. This new approach is part of the Copilot Cowork initiative, specifically targeting the Researcher tool within Microsoft 365 Copilot.

The standout feature here is called 'Critique.' Instead of relying on a single AI model, Researcher will now use GPT to generate an initial response, which is then meticulously refined and critiqued by Claude. Microsoft suggests that this dual-model architecture creates a high-quality feedback loop. It seems this method is designed to mimic professional and academic research settings, where peer review is essential for factual accuracy and analytical depth.

For those who want a bit more transparency in how their AI thinks, Microsoft is also introducing the 'Model Council' feature. This allows users to see side-by-side responses from both Anthropic and OpenAI models. It even provides a detailed report highlighting where the two AI giants agree or disagree on a topic. It’s like having two expert consultants on your screen at once, letting you decide which perspective fits your needs best.

Currently, these advanced features are available through the Microsoft 365 Copilot Frontier program. This acts as an early-access sandbox for the company's most experimental AI innovations. Early benchmarks suggest that this multi-model approach is already outperforming some of the latest deep research models in terms of completeness and objectivity. It looks like the era of 'one AI to rule them all' might be evolving into a more collaborative ecosystem.