AI Research Atlas

MedGemma 27B Multimodal and MedSigLIP

Google · 9 July 2025

MedGemma 27B Multimodal adds image and longitudinal health-record understanding to the open MedGemma family; MedSigLIP is a lightweight medical image-text encoder.

Complements the earlier 4B multimodal and 27B text-only MedGemma models with multimodal 27B and longitudinal health-record interpretation. Pitched as efficient, privacy-preserving starting points for health developers. Weights on Hugging Face.

Date
Wednesday, 9 July 2025
Lab
Google
Kind
open-weights
Access
open weights (restricted license)

This post (2025-07-09) announces the 27B multimodal and MedSigLIP; the original MedGemma collection predates it.

Sources

  1. research.google/blog/medgemma-our-most-capable-open-models-for-health-ai-development/

This record was checked against its sources on 6 October 2026. How we check

Related