Microsoft study reports AI Diagnostic Orchestrator outperforms doctors on historical medical cases

What happened

Microsoft's GPT-5-powered Diagnostic Orchestrator outperformed physicians on historical medical cases in research published Thursday in Science. Its multiple AI agents achieved around 83 percent accuracy on cases from The New England Journal of Medicine, versus roughly 36 percent for generalist physicians using references and 22 percent without them. The system still needs clinical validation, and Microsoft plans hospital testing with Beth Israel Deaconess Medical Center, likely starting in the next couple of months.

Read next

• More changes in today’s briefingMicrosoft Research and Collaborator Resolve the KLS Conjecture With Human-Directed AI

The story so far

Microsoft reported in Science that its GPT-5-powered Diagnostic Orchestrator outperformed physicians on historical medical cases, while validation in clinical settings remains pending. On cases drawn from The New England Journal of Medicine, the system achieved around 83 percent diagnostic accuracy, compared with roughly 36 percent for generalist physicians using references and 22 percent for those without them. The system uses multiple AI agents to assess symptoms and medical histories, choose tests and arrive at a diagnosis. The research evaluated scenarios drawn from medical case reports and emergency department data; those results do not establish how well it performs with patients in clinical settings. Microsoft plans to test the system in hospitals with Beth Israel Deaconess Medical Center, likely beginning in the next couple of months.