Microsoft published results in Science on October 22 showing its AI diagnostic system, MAI-DxO, correctly diagnosed complex medical cases 83% of the time in simulation, compared with 36% for primary care doctors and general internists working individually with reference materials2.
The system, which Microsoft calls Diagnostic Orchestrator, coordinates multiple AI agents that act like a panel of physicians: questioning a simulated patient, ordering tests, and debating a diagnosis. Microsoft says the tool, powered by GPT-5, can diagnose conditions up to four times more accurately than human doctors1,3.
The simulation drew on complex cases from the New England Journal of Medicine, compiled into a benchmark Microsoft calls SDBench, which it plans to open-source. The human comparators were generalist physicians working alone; specialists and physician teams, who would typically handle cases of this complexity, were not part of the comparison.
The study's own data indicates that much of the performance gain comes from more powerful underlying AI models, with Microsoft's orchestration layer adding incremental improvement on top.
Mustafa Suleyman, Microsoft AI's CEO, called the result a step toward "medical superintelligence". In an interview with The Verge, Suleyman said: "We're going to build something that is indistinguishable from human performance … in the next couple of years".
Dominic King, vice president of Health at Microsoft AI, told TechCrunch that "the biggest challenge in healthcare delivery at the moment is the growing mismatch between demand and supply of medical expertise".
Microsoft joins Google and OpenAI in claiming its AI can match or outperform doctors, according to TechCrunch. ANALYSIS The 83%-versus-36% gap is large, but the absence of specialist physicians from the comparison limits the clinical conclusions that can be drawn from the study; the benchmark measures the system against a lower baseline than the teams that typically manage such cases.