Microsoft AI Correctly Diagnoses 83% of Complex Cases in Science Study

For patients facing a baffling illness, the standard path often involves referrals, waiting, and the uncertainty of whether the right specialist has been found. A study published on Thursday, October 22, 2026, suggests a future where that delay might be unnecessary. According to research published in the journal Science, an artificial intelligence system developed by Microsoft demonstrated a significantly higher ability to correctly diagnose complex medical cases than physicians working individually.

The central finding of the research is a stark contrast in diagnostic accuracy. The system, known as MAI-DxO, correctly identified the conditions in 83% of the complex cases presented to it. By comparison, primary care doctors and general internists, working alone but with access to reference materials, arrived at the correct conclusion in 36% of instances. Mustafa Suleyman, the CEO of Microsoft AI, described the findings as a step toward 'medical superintelligence.'

The system operates by coordinating several AI agents, allowing them to function like a panel of doctors who question a patient, order tests, and debate a diagnosis. The cases used to test this capability were complex problems typically handled by specialists or physician teams, who were not part of the direct comparison. Dominic King, vice president of Health at Microsoft AI, stated, 'There is something very substantial and exciting here,' adding that the biggest challenge in healthcare delivery is the growing mismatch between demand and supply of medical expertise.

The study's own data indicates that much of the improvement stems from more powerful underlying AI models, with Microsoft's orchestration system providing additional gains on top of that capability. The results, however, are not uniform across all types of cases. The system's advantages were most pronounced when tackling complex problems and when assisting less capable AI models. For instance, Gemini 2.5 Flash scored 46.4% on its own but nearly 69% with the assist from DxO. Even so, the gap between the weaker and stronger models was larger than the boost DxO provided.

Further details reveal limitations. On cases typically seen in emergency departments, the system provided little improvement in accuracy or cost compared to using a standard large language model on its own. Microsoft's researchers said that accuracy gains naturally shrink as models approach the theoretical maximum. 'The DxO orchestration helps with the broader challenge of optimization both accuracy and cost,' a spokesperson added.

Looking ahead, Microsoft is expanding its efforts. The company is working with the Mayo Clinic to train a foundation model for healthcare, with release planned for next year. Suleyman believes this model will enable the company to 'make really enormous leaps here,' and says DxO is part of the foundational infrastructure Microsoft is building to see real breakthroughs in healthcare.

To move beyond simulations, Microsoft plans to evaluate the clinical reasoning capabilities of DxO across different hospital settings through a new partnership with Beth Israel Deaconess Medical Center in Boston. Patients will receive standard treatment, including triage and time with a specialist, while those who consent will also engage with DxO. The company will compare the recommendations made by the system against those of human doctors.

Suleyman has spent his career trying to use AI to solve medical issues, previously co-founding DeepMind, which was acquired by Google in 2014. The following year, Google DeepMind acquired King’s startup Hark, a clinical task management app, and the two of them launched DeepMind Health together. Suleyman joined Microsoft to lead its AI unit in 2024, and King followed a few months later. He stated that the use of generative AI in healthcare is still early, but expects healthcare models to get 'an order of magnitude better in capability within the next 24 to 36 months.'

Through a new partnership with Beth Israel Deaconess Medical Center in Boston, Microsoft will evaluate the clinical reasoning capabilities of DxO across different hospital settings. Patients will receive standard treatment, including triage and time with a specialist, while those who consent will also engage with DxO. The company will compare the recommendations made by the system against those of human doctors.