Google has released a suite of open-source AI models, dubbed MedGemma, designed to accelerate healthcare innovation. The release includes MedGemma 27B Multimodal, capable of processing both medical text and images, and MedSigLIP, a model trained to connect visual information with corresponding textual descriptions in medical contexts. A smaller, but powerful model, MedGemma 4B, has also been made available, showcasing high accuracy in analyzing chest X-ray reports, as validated by radiologists. These models are intended to be leveraged by hospitals, researchers, and developers. Google stresses that MedGemma models are tools to assist medical professionals and require human oversight to ensure responsible application.