Key Findings
In a study published in Science, Microsoft’s AI diagnostic system, MAI-DxO (DxO), correctly diagnosed complex medical cases from the New England Journal of Medicine 83% of the time. In comparison, primary care doctors and general internists working individually with reference materials were correct 36% of the time. However, the study noted that these complex cases are typically handled by specialists or physician teams, who were not included in the comparison.
Technical Details
DxO functions as an orchestration system that coordinates several AI agents to act as a panel of doctors by questioning patients, ordering tests, and debating diagnoses. Microsoft tested the system using models from Meta, Anthropic, Google, DeepSeek, and OpenAI. The best results were achieved with GPT-5 and GPT-5.5.
- GPT-5: Scored 78.9% accuracy on complex cases alone at approximately $8,400 per case; DxO reduced the cost by about $3,000 while increasing accuracy by roughly four points.
- GPT-5.5: DxO increased accuracy from 72% to nearly 87% for complex cases at a slightly lower cost.
- Gemini 2.5 Flash: Accuracy rose from 46.4% to nearly 69% when using DxO.
The system showed little to no improvement in accuracy or cost when applied to emergency department cases compared to standalone models.
Background and Context
Mustafa Suleyman, CEO of Microsoft AI, and Dominic King, VP of Health, are leading the effort to address the mismatch between medical expertise supply and demand. Microsoft is currently partnering with the Mayo Clinic to train a healthcare-specific foundation model for release next year. Additionally, a partnership with Beth Israel Deaconess Medical Center in Boston will evaluate DxO’s clinical reasoning in real-world hospital settings, comparing its recommendations against human doctors in parallel with standard treatment.
Significance and Outlook
Suleyman predicts a “coding style breakthrough” for medical records by 2027, enabling highly accurate predictions of how records unfold. He expects healthcare model capabilities to improve by an order of magnitude within 24 to 36 months as the company implements a “flywheel” of real-world usage and iteration.
Get our weekly technology intelligence — free
Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.
Subscribe Free — Weekly Tech Intelligence
By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.
- Your email and selected fields are used only to deliver the newsletter.
- We never share your information with third parties.
- You can unsubscribe anytime via the link in each email.
See our Privacy Policy for details.
Takes about a minute · Unsubscribe anytime
