How accurate was Microsoft's AI compared with human doctors?

On cases from The New England Journal of Medicine, Microsoft's system achieved about 83 percent diagnostic accuracy versus roughly 36 percent for generalist physicians using references and 22 percent for those working without them.

What powers Microsoft's Diagnostic Orchestrator?

The system is powered by OpenAI's GPT-5 and uses multiple AI agents to reason through symptoms and medical histories, decide what tests to order, and reach a diagnosis.

Has Microsoft tested the system on real patients yet?

No. The system has only been tested on past medical cases, and Microsoft plans to begin testing it in hospital settings with Beth Israel Deaconess Medical Center in the next couple of months.

Large Language ModelsAI in HealthcareThe Verge AI4m agoOct 23, 2026, 04:00 JST

Microsoft's Diagnostic Orchestrator beats doctors in Science study

Send on LINE

3 Key Points

What happened
Microsoft's Diagnostic Orchestrator, powered by OpenAI's GPT-5, scored about 83 percent diagnostic accuracy on New England Journal of Medicine cases, versus roughly 36 percent for generalist physicians, per findings in Science.
Why it matters
The results, which Microsoft says show AI diagnosing conditions up to four times more accurately than human doctors, are based only on past cases and still need clinical validation, CEO Mustafa Suleyman says.

WHO IT HITSHospital systems and clinicians evaluating AI-assisted diagnosis will face early evidence that a multi-agent system can outperform physicians on archived cases, but adoption likely hinges on the real-world clinical testing Microsoft says is still ahead.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

Microsoft's Diagnostic Orchestrator was first developed when GPT-4.1 was the most advanced model available, and the researchers say subsequent testing suggests that improvements in model capability do not necessarily diminish the benefits of having multiple agents work together. The peer-reviewed paper does not include newer models such as GPT-6 Astra, Meta's Muse Wonder Full, Google's Gemini 4 Argon, or Anthropic's Opus 5.5.

The research is part of a broader ambition at Microsoft to develop what the company calls medical superintelligence, a goal it says would outperform the best human-led healthcare systems. Diagnosis is only one area; the company has outlined five others AI will need to master, including coordinating clinical care and making treatment decisions. The path to deployment is constrained by regulation and by the need for real-world evidence, which Microsoft acknowledges is still missing.

FAQ

How accurate was Microsoft's AI compared with human doctors?: On cases from The New England Journal of Medicine, Microsoft's system achieved about 83 percent diagnostic accuracy versus roughly 36 percent for generalist physicians using references and 22 percent for those working without them.
What powers Microsoft's Diagnostic Orchestrator?: The system is powered by OpenAI's GPT-5 and uses multiple AI agents to reason through symptoms and medical histories, decide what tests to order, and reach a diagnosis.
Has Microsoft tested the system on real patients yet?: No. The system has only been tested on past medical cases, and Microsoft plans to begin testing it in hospital settings with Beth Israel Deaconess Medical Center in the next couple of months.

The Verge AIRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.