Microsoft AI diagnostic system outperforms doctors on complex cases
The MAI-DxO system achieved 83% accuracy on complex medical cases in a study published in Science, compared to 36% for generalist physicians using reference materials.

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 Science. In simulations using complex cases from the New England Journal of Medicine, the system was correct 83.2% of the time, while primary care doctors and general internists using reference materials were correct 36.1% of the time.
The system coordinates several AI agents that act like a panel of doctors to reason through symptoms, medical histories, and test orders. It is a model-agnostic system that simulates a panel of clinical experts with complementary roles and applies cost-conscious test stewardship. While the system performed well on complex cases, it showed little improvement in accuracy or cost on emergency department cases when compared to using GPT-5 on its own.
Mustafa Suleyman, CEO of Microsoft AI, called the result a step toward medical superintelligence. He said the company is working with the Mayo Clinic to train a healthcare foundation model, with a release planned for next year. Suleyman expects healthcare models to get an order of magnitude better in capability within the next 24 to 36 months.
The study's data suggests much of the improvement comes from more powerful underlying AI models, with Microsoft’s orchestration system adding gains on top. The system was tested with models from several companies, including OpenAI, Google, Anthropic, Meta, and DeepSeek, but its best results came with OpenAI’s GPT-5 and GPT-5.5. On complex cases, the DxO system increased GPT-5.5’s accuracy from 72% to nearly 87% at a slightly cheaper cost.
Microsoft will evaluate the clinical reasoning capabilities of DxO across different hospital settings through a new partnership with Beth Israel Deaconess Medical Center in Boston. The company and its partners will compare DxO’s recommendations against those of human doctors.
Key facts
- MAI-DxO achieved 83.2% accuracy on complex New England Journal of Medicine cases.
- Generalist physicians using reference materials achieved 36.1% accuracy in the same study.
- Microsoft is working with the Mayo Clinic to train a healthcare foundation model for release next year.
Three perspectives
Neutral
Microsoft's MAI-DxO system achieved 83% accuracy on complex medical cases in a simulation, outperforming generalist physicians who scored 36%.
Positive
Mustafa Suleyman said the results are a step toward medical superintelligence and that the company is working with the Mayo Clinic to train a healthcare foundation model.
Negative
The system showed little improvement in accuracy or cost on emergency department cases compared to using GPT-5 alone.
Your week in news, every Wednesday
The biggest stories of the week for your edition, in your inbox every Wednesday morning, your time. Free, and you can unsubscribe at any time.
