Benchmark

SlakeVQA

A semantically-labeled knowledge-enhanced dataset for medical visual question answering. Contains 642 radiology images (CT scans, MRI scans, X-rays)…

Modality
multimodal
Categories
multimodal, reasoning, image_to_text, healthcare, vision
Openness
unknown
Source
llm_stats
Reported scores
4

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
MedGemma 4B ITGoogle0.6232025-05-20evidence
Qwen3.5-122B-A10BQwen0.8162026-02-24evidence
Qwen3.5-27BQwen0.82026-02-24evidence
Qwen3.5-35B-A3BQwen0.7872026-02-24evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.