Gemma is a family of generative artificial intelligence (ai) models and you can use them in a wide variety of generation tasks, including question answering, summarization, and reasoning. Whether you're a developer testing ideas or a business deploying solutions, we make advanced ai accessible without the need for massive computing resources. This repository contains the implementation of the gemma pypi package. Gemma is a lightweight, family of models from google built on gemini technology The gemma 3 models are multimodal—processing text and images—and feature a 128k context window with support for over 140 languages. It is based on similar technologies as gemini
The first version was released in february 2024, followed by gemma 2 in june 2024 and gemma 3 in march 2025. Gemma 3n was created in close collaboration with leading mobile hardware manufacturers Engineered for speed and quality, with a significantly reduced memory footprint. This page documents releases for the gemma family of models Release of vaultgemma in 1b parameter size Release of embeddinggemma in 308m parameter size
Release of t5gemma across different parameter sizes Release of medgemma 27b parameter multimodal model Release of gemma 3n in e2b and e4b sizes.
OPEN
