ReXVQA: leaderboard
ReXVQA (Harvard Rajpurkar Lab, 2025) is a chest X-ray visual question answering benchmark run as ReXrank Challenge V2.0, with multiple-choice questions about radiographs. The score is overall test-set accuracy (%); higher is better.
Metric: Accuracy (%). Source: rexrank.ai. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MedGemma-4B-IT | 83.44 | #2732 |
| 2 | Qwen 2.5 VL 7B Instruct | 68.67 | #1842 |
| 3 | Qwen 2 VL 7B Instruct | 68.35 | #2569 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=rexvqa · How It Works · Data refreshed daily, snapshot 2026-11-12.