Microsoft AI Suleyman Claims Diagnostic System Outperforms General Physicians

A Microsoft AI system demonstrated higher accuracy than generalist doctors in solving difficult medical puzzles in a simulated environment. The company released the findings on October 22, 2026, outlining a specific strategy to develop tools capable of superhuman-level medical reasoning.

The system, known as MAI-DxO, functions by coordinating multiple AI agents that simulate a panel of physicians. These agents are designed to question a patient, order necessary tests, and debate potential diagnoses. In a simulation involving complex cases drawn from the New England Journal of Medicine, MAI-DxO reached a correct diagnosis 83% of the time. By comparison, primary care doctors and general internists working individually with reference materials achieved a 36% accuracy rate.

Mustafa Suleyman, CEO of Microsoft AI, described the outcome as a step toward 'medical superintelligence.' Dominic King, vice president of Health at Microsoft AI, noted, 'There is something very substantial and exciting here.' He added, 'The biggest challenge in healthcare delivery at the moment is the growing mismatch between demand and supply of medical expertise.'

However, the study includes significant caveats regarding the comparison. The complex cases used in the simulation are typically managed by specialists or physician teams, neither of whom were part of the comparison group. Furthermore, the study's own data indicates that much of the performance gain stems from the underlying AI models rather than the orchestration system itself.

Microsoft tested MAI-DxO using models from several companies, including OpenAI, Google, Anthropic, Meta, and DeepSeek. The highest accuracy scores were achieved using OpenAI’s GPT-5 and GPT-5.5. Regarding the cost of running these AI systems, GPT-5 on its own scored 78.9% on complex cases at approximately $8,400 per case. While MAI-DxO reduced costs by about $3,000, it improved accuracy by only about four points.

Results varied when using GPT-5.5, a model released in April 2026. On complex cases, MAI-DxO raised GPT-5.5’s accuracy from 72% to nearly 87% at a slightly lower cost. However, the system did not show significant improvement on emergency department cases. For those scenarios, both the raw model and the system achieved approximately 85% accuracy at a similar cost.

Microsoft is now moving toward real-world testing through a partnership with Beth Israel Deaconess Medical Center in Boston. Patients will receive standard treatment, including triage, specialist consultations, and investigations, while consenting participants will also engage with MAI-DxO. Microsoft and its partners plan to compare the system's recommendations against those of human doctors.

This initiative coincides with a manifesto regarding the construction of medical superintelligence, which Microsoft defines as 'the capacity to surpass the strongest contemporary human-led health systems in medical reasoning and coordination across the full arc of care.' Suleyman expressed confidence in the company's trajectory, stating, 'We have to have the absolute number one state-of-the-art medical clinical decision support products over the next few years.'