RadM-Bench

RadM-Bench: A Bilingual Multimodal Benchmark for Diagnostic Radiology 📖 Overview RadM-Bench is a bilingual (English–Chinese) multimodal benchmark for evaluating the diagnostic performance of multimodal large language models (MLLMs) in radiology. It is built to expose three blind spots in existing benchmarks: the gap between curated 2D snapshots and real volumetric (3D) imaging, the gap between public teaching cases and routine clinical practice, and the gap… See the full description on the dataset page:

Source: Hugging Face Hub (Elizabeth123/RadM-Bench). Metadata imported from the dataset’s Hub tags.