At Google I/O 2025, Google launched Gemma 3n, an AI model designed for efficient performance on various devices without needing cloud support. The model can process audio, text, images, and videos, addressing privacy concerns prevalent in AI development. In addition, the unveiling of MedGemma aims to enhance health-related analytics, while SignGemma will facilitate communication for deaf and hard-of-hearing individuals. Despite facing criticism regarding licensing terms, the popularity of Gemma models among developers persists as downloads reach into the tens of millions.
Gemma 3n is designed to run smoothly on phones, laptops, and tablets, offering capabilities in audio, text, images, and videos while emphasizing privacy.
Google is releasing MedGemma, its most capable model for analyzing health-related text and images, under the Health AI Developer Foundations program.
Despite criticism regarding Gemma's non-standard licensing terms, developers have collectively downloaded the models tens of millions of times, indicating strong interest.
With the upcoming SignGemma, Google promises to enable apps that help deaf and hard-of-hearing users through the translation of sign language into text.
Collection
[
|
...
]