Medical AI: a strong retrospective benchmark
Microsoft reports 85.5% accuracy for the best MAI-DxO/o3 configuration on its published NEJM case benchmark.
Source finding
Microsoft reports 85.5% accuracy for the best MAI-DxO/o3 configuration on its published NEJM case benchmark.
Our analysis
It shows a promising research direction for diagnostic reasoning.
What remains uncertain
The 2025 result is not a prospective clinical trial or proof of replacing doctors.