Microsoft's AI diagnostic system MAI-DxO is 1.7x more accurate than doctors

Microsoft says its AI diagnostic system MAI-DxO reached 1.7 times the accuracy of human doctors on a dataset of common US emergency room cases. The findings were published in the journal Science.

The system runs multiple AI agents that work like a panel of doctors. They ask patients questions, order lab tests, and debate possible diagnoses together.

On complex cases from the New England Journal of Medicine, MAI-DxO scored 83% accuracy, while family doctors with access to reference materials got only 36%. In real life, such cases usually go to specialists, who were not part of the test. The system costs about half to a third of what doctors cost.

Microsoft says the gains come mainly from stronger base AI models, with its orchestration adding extra boosts. After testing models from OpenAI, Google, Anthropic, Meta and DeepSeek, OpenAI's GPT-5 and GPT-5.5 performed best.