RadMultiBench: A Reproducible Dual-Task Benchmark Exposing the Limits of Pre-Trained Vision-Language Models Across Thoracic and Breast Imaging
Md Nahin Alam, Md Tanjil Rahman, Devprosad Sarkar, Ahmed Faizul Haque Dhrubo, Mohammad Abdul Qayum, Mohammad Ashrafuzzaman Khan
RadMultiBench evaluates pre-trained vision-language models on medical imaging tasks, showing performance gaps in report generation and classification across thoracic and breast imaging.
bioRxiv · Nov 6, 2026 · PreprintBioinformatics