Microsoft AI System Outperforms Doctors on Complex Medical Cases in Simulation
Microsoft's MAI-DxO AI system correctly diagnosed 83% of complex cases in a simulation, versus 36% for primary care doctors and general internists. The company calls it a step toward "medical superintelligence," though the study's data suggests much of the gain comes from stronger underlying models.
Microsoft's AI diagnostic system, MAI-DxO, correctly diagnosed complex medical cases in a simulation more than twice as often as generalist physicians, according to a study published Thursday in the journal Science. Mustafa Suleyman, CEO of Microsoft AI, described the result as a step toward "medical superintelligence."
The system coordinates several AI agents that work like a panel of doctors: they question a patient, order tests and debate a diagnosis. In a simulation using complex cases from the New England Journal of Medicine, it was right 83% of the time. Primary care doctors and general internists, working individually with reference materials, were right 36% of the time. Cases this complex are normally handled by specialists or physician teams, who were not part of the comparison.
What drives the improvement
The study's data suggests much of the improvement comes from more powerful underlying AI models, with Microsoft's orchestration layer adding gains on top. The system was tested with models from several companies, and results were best with OpenAI's GPT-5 and GPT-5.5. GPT-5 alone scored 78.9% on complex cases at about $8,400 per case; with MAI-DxO the cost fell by about $3,000 while accuracy rose only around four points. GPT-5.5's accuracy rose from 72% to nearly 87%. On emergency department cases there was little improvement. The system helped most with weaker models: Gemini 2.5 Flash went from 46.4% to nearly 69%.
Next steps
Microsoft is working with Mayo Clinic to train a healthcare foundation model, planned for release next year. Through a partnership with Boston's Beth Israel Deaconess Medical Center, MAI-DxO will be evaluated in real hospital settings: patients will receive standard care, while those who consent will also interact with the system, and its recommendations will be compared with those of human doctors. Google published its own clinical study of its AMIE diagnostic AI in The Lancet this month.
Suleyman expects healthcare models to become an order of magnitude better in capability within 24 to 36 months.
