Google's MedGemma open medical VLM lands in Nature Medicine

Google's MedGemma, a Gemma 3-derived family of medically tuned vision-language models, reports 86.2% accuracy on MedQA in a Nature Medicine paper that describes it as the best open model performance at release.

The release spans four artifacts: a 4-billion-parameter multimodal model, a 27B multimodal model, a 27B text-only variant, and MedSigLIP, a 400-million-parameter medical vision encoder. Against the base Gemma 3 models, the authors report gains of 15.5–18.1% in chest X-ray finding classification, 2.6–10% in medical image question answering, and 10.8% in agentic evaluations.

On radiology report generation using MIMIC-CXR chest X-rays, the paper reports that "81% of generated reports were rated likely to lead to the same or superior clinical decisions," with 68% of normal and 49% of abnormal studies judged equal or better than human-written originals.

The authors frame the family as "exceeding the performance of similarly sized generative models while maintaining the general capabilities of the Gemma base models." Two of the researchers we track pushed the paper into their feeds the week it published.