Search
On January 14, 2026, Google introduced its latest open-source medical AI model, MedGemma 1.5 4B, along with a specialized speech recognition model, MedASR. MedGemma 1.5 4B stands out as a lightweight yet powerful model that can be deployed locally. It excels in analyzing not only textual medical data but also three-dimensional imaging data, such as CT and MRI scans, offering improved accuracy compared to its previous version. Meanwhile, MedASR is designed specifically for accurate speech recognition of medical terminology, boasting a low transcription error rate and effortless integration with the MedGemma family of models. Google has opted to make the MedGemma series available as open-source, providing developers with the tools to create innovative medical AI applications.
-
C114 Communication Network -
Communication Home
