The "gatekeeper mentality" of the AI era is accelerating its decline, and even tech giant Microsoft is beginning to enhance its product capabilities by integrating competitors' strengths.

On March 31, according to media reports, Microsoft announced a major update to its 365 Copilot research agent function. The core of this update is that users can now simultaneously call both OpenAI's GPT model and Anthropic's Claude model within the same research task, allowing these two top-tier large models to "work side by side."

image.png

New Collaboration Paradigm: GPT as the Lead Writer, Claude as the Reviewer

Microsoft has introduced a new feature called "Critique (Criticism)," which sets up a rigorous workflow for in-depth research:

Clear Division of Labor: The system first generates a draft using OpenAI's GPT, then hands it over to Anthropic's Claude to review following an academic evaluation process.

Quality Control: Claude will focus on identifying flaws and improving the report's accuracy, completeness, and citation quality.

Two-Way Simulation: Microsoft stated that this feature will support bidirectional operation in the future, meaning Claude could write first and GPT review afterward.

image.png

"Council" Mechanism: Let Models Act as Judges

Aside from collaborative work, Microsoft also launched a synchronized research mechanism called "Council (Council):"

Independent Research: Let GPT and Claude independently conduct research on the same topic and generate reports.

Judgment Summary: Finally, a "judgment model" evaluates the two reports, extracting common points and objectively pointing out existing differences and each model's unique contributions.

Data Support: 1+1 Is Indeed Greater Than 2

According to the DRACO benchmark test, which measures the quality of deep research, the output from multi-model collaboration clearly outperforms that of a single model. Nicole Herskowitz, Vice President of Microsoft 365 and Copilot, stated that this approach effectively limits AI hallucinations and significantly improves the reliability of productivity.

Industry Background: "Alliances Among Giants"

Although Microsoft has a deep relationship with OpenAI, recent actions show flexibility in its strategy:

Breaking Monopolies: After last year's restructuring agreement, Microsoft gained the freedom to collaborate with OpenAI's competitors.

Heavy Investment: Microsoft has previously announced strategic partnerships with NVIDIA and Anthropic, and invested up to $5 billion in Anthropic.

Market Reaction: As of the time of writing, Microsoft's stock price rose nearly 2%, although it is still somewhat below its peak from last year. However, the continuous integration of AI features is seen as a key factor supporting its fundamentals.

Conclusion: From "Sole Combat" to "Intelligent Swarms"

When GPT's creativity meets Claude's rigor, Microsoft is pushing the practicality of AI agents to a new height through a "computing power combination." For users, there's no need to worry about choosing which model to use, because the strongest AI assistant has already achieved "cross-factory collaboration" at the underlying level.