Microsoft's MAI-DxO solves 83.2% of complex cases vs physicians' 36.1%
emmanuelvivier · x · 2026-10-25
Microsoft's MAI Health team published official results for MAI-DxO, its AI diagnostic system:
- On 304 complex cases from the New England Journal of Medicine, MAI-DxO solved 83.2%, versus 36.1% for physicians with access to textbooks and the internet — nearly 4x better when doctors had no aids
- On common US emergency-department presentations (MIMIC-IV-ED benchmark), it was 1.7x more accurate than physicians
- It reached correct diagnoses at half to one-third the cost, using fewer tests
Microsoft frames it as a step toward "medical superintelligence," positioning the system as a step-by-step clinical diagnostic assistant.
More from Models
- Pro user: 62,500 Codex credits ($2,500) lock to smallest models the day your plan ends — lanqo88 · 2026-10-25
- Could OpenAI throttle Codex usage differently per user? Redditors demand transparency — Dadideology · 2026-10-25
- Claude notices it's Sunday for a European user and tells them to go touch grass — aiamblichus · 2026-10-25
- Claude Models Now GA on Google Cloud with New Anthropic-Managed Deployment Option — emmanuelvivier · 2026-10-25
- GPT-6 Astra Hits 39% on OpenDS-World Live Web Deep Search Benchmark vs 21.1% Human Average — emmanuelvivier · 2026-10-25
- Study: Grok picks Republicans 92%, GPT-6.1 Sol picks Democrats 97% — emmanuelvivier · 2026-10-25