Alibaba DAMO Academy

Healthcare AI · Medical Imaging · China Tech

Milestone

Alibaba's DAMO RADAR AI outperforms 23 of 26 radiologists on 146 diseases

Free access to a single scan-reading system undercuts an industry built on per-disease detection fees for narrow AI tools.

• Alibaba's DAMO Academy published a peer-reviewed study in Science on Sept. 18, 2026 unveiling DAMO RADAR, a generalist AI that reads abdominal CT scans for 146 diseases, and released its code and weights for free.
• RADAR is a vision-language model that reads a single contrast-enhanced CT covering 18 abdominal organs, replacing radiology AI's usual approach of one narrow model per disease.
• Trained on 424,911 exams with 1.5 million image-text pairs, RADAR scored a mean AUC of 0.913 across 146 findings, versus 0.776 for the best competing vision-language model.
• In a study with 26 radiologists from multiple hospitals, RADAR outperformed 23 of them on average, and as a diagnostic aid it cut missed diagnoses by 10% and diagnosis time by more than 30%.
• The release challenges radiology AI's decade-long one-disease-one-model paradigm, arriving as researchers project a shortfall of 750,000 radiologists worldwide by 2050.
• The code is Apache 2.0-licensed but the model weights carry a CC BY-NC-SA 4.0 license barring commercial deployment, and RADAR has not yet received U.S. FDA clearance.
• By giving away a system that already beats most specialists, Alibaba is testing whether free, open generalist AI can outcompete the paid, narrow-detector market that dominates radiology today.