Microsoft’s MAI-DxO AI system diagnosed complex medical cases correctly 83% of the time in a Science study. Generalist doctors scored 36% on the same cases. The system uses multiple AI agents to question patients, order tests, and debate diagnoses. Most gains came from the underlying AI models rather than Microsoft’s orchestration. GPT-5 alone cost about $8,400 per case and scored 78.9% accuracy. DxO cut that cost by $3,000 but only raised accuracy by four points. GPT-5.5 with DxO reached nearly 87% accuracy on complex cases. Microsoft is training a healthcare foundation model with Mayo Clinic for release next year. Suleyman called this a step toward medical superintelligence. A real-world study with Beth Israel Deaconess Medical Center will test DxO in hospital settings. Microsoft defines medical superintelligence as surpassing human-led health systems in reasoning and coordination. The figures come from a Microsoft study.

• 83% correct diagnoses by MAI-DxO
• 36% correct by generalist doctors
• $8,400 per case for GPT-5 alone
• $3,000 cost reduction with DxO