VQA-Rad
multimodal
+
+
+
+
About
VQA-RAD is the first manually constructed medical visual question answering benchmark featuring 315 radiology images with 3,515 clinician-generated question-answer pairs. This specialized evaluation tests AI models' ability to analyze radiological images and answer medically relevant questions, combining medical knowledge with visual understanding to assess capabilities in clinical image interpretation and medical reasoning.
+
+
+
+
Evaluation Stats
Total Models1
Organizations1
Verified Results0
Self-Reported1
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
1 models
Top Score
49.9%
Average Score
49.9%
High Performers (80%+)
0Top Organizations
#1Google
1 model
49.9%
+
+
+
+
Leaderboard
1 models ranked by performance on VQA-Rad
| License | Links | ||||
|---|---|---|---|---|---|
| May 20, 2025 | Health AI Developer Foundations terms of use | 49.9% |
