Microsoft’s diagnostic agents beat generalist doctors in a Science study
On complex published cases MAI-DxO was right 83% of the time against 36% for generalist physicians working alone, though much of the gain comes from the underlying models.
Source published · In the Daily Brief of 23 October 2026
Filed under Agents, Large language models
Microsoft AI’s diagnostic orchestrator coordinates several agents that question a simulated patient, order tests and debate a diagnosis. On complex cases from the New England Journal of Medicine it reached the correct diagnosis 83% of the time, according to the study published in Science on 22 October. Primary care doctors and general internists, working individually with reference material, scored 36%. Specialists, who usually handle such cases, were not part of the comparison.
The orchestration adds most to weaker models and to hard cases. Gemini 2.5 Flash rose from 46.4% to nearly 69%, and GPT-5.5 from 72% to nearly 87% on complex cases, while on emergency-department cases accuracy stayed around 85% with or without it. The best results came with OpenAI’s GPT-5 and GPT-5.5.