Google's Gemma AI model family gets new addition: What's special?
21 May 2025

At the I/O 2025 event, tech giant Google unveiled the latest member of its Gemma family of AI models - the Gemma 3n.

The new 'open' model is capable of working seamlessly across smartphones, laptops, and tablets.

As Google announced at the event, Gemma 3n has been designed to handle audio, text as well as images and videos with ease.

Shift toward offline AI models
Offline models

In recent years, the trend of building efficient AI models capable of running offline without cloud computing support has gained a lot of attention.

Not only are these models cheaper than their larger counterparts, but they also provide better privacy by eliminating the need to transfer data to a remote data center.

This shift marks an important development in artificial intelligence, efficiency, and user privacy.

Google launches MedGemma and SignGemma
New models

Along with Gemma 3n, Google is also launching MedGemma via its Health AI Developer Foundations program.

MedGemma, as the company says, is its most advanced open model for analyzing health-related text and images.

Additionally, Google has announced plans for SignGemma - an open model to translate sign language into spoken-language text.

This innovative tool would help developers create new apps and integrations for deaf and hard-of-hearing users.

Gemma models face criticism despite popularity
Licensing concerns

Despite criticism over its custom, non-standard licensing terms (which some developers argue make commercial use of the models a risky venture), Gemma has been downloaded tens of millions of times, collectively.

The widespread adoption shows that despite concerns over licensing terms, there is a lot of interest and demand for these AI models in the developer community.