Google’s Gemma AI Model Family Expands with Gemma 3n, MedGemma, and SignGemma

Google has unveiled Gemma 3n, the latest addition to its family of “open” AI models, during the Google I/O 2025 event on Tuesday. Gemma 3n is designed to run smoothly on various devices, including phones, laptops, and tablets. The model is capable of handling audio, text, images, and videos, according to the tech giant.

Offline AI models, which do not require cloud computing, have gained popularity in the AI community in recent years. These models are not only cost-effective but also preserve user privacy by eliminating the need to transfer data to remote data centers. Gemma Product Manager Gus Martins announced that Gemma 3n can run on devices with less than 2GB of RAM. He added that Gemma 3n shares the same architecture as Gemini Nano and is engineered for incredible performance.

In addition to Gemma 3n, Google is releasing MedGemma through its Health AI Developer Foundations program. MedGemma is Google’s most capable open model for analyzing health-related text and images. Martins explained that MedGemma is a collection of open models for multimodal health text and image understanding. The model works well across a range of image and text applications, allowing developers to adapt the models for their health apps.

Another upcoming addition to the Gemma family is SignGemma, an open model designed to translate sign language into spoken-language text. Google claims that SignGemma will enable developers to create new apps and integrations for deaf and hard-of-hearing users. Martins said, “SignGemma is a new family of models trained to translate sign language to spoken-language text, but it’s best at American Sign Language and English. It’s the most capable sign language understanding model ever, and we can’t wait for you — developers and deaf and hard-of-hearing communities — to take this foundation and build with it.”

It is worth noting that Gemma has faced criticism for its custom, non-standard licensing terms, which some developers argue make using the models commercially risky. Despite this, developers have downloaded Gemma models tens of millions of times collectively.