Clément Farabet

video input now rolling out for chatgpt voice mode including screenshare! (and santa mode for a little holiday treat)

200 million times! 🚀🚀 That’s how many times developers, researchers, and startups have downloaded the Gemma 4 family in just 2.5 months. When we built Gemma, our goal was to bring Google’s Gemini research to the open-science community in an accessible and highly capable package. Seeing this milestone today is both humbling and exhilarating. From lightweight agents to satellites orbiting Earth, the community is pushing the boundaries of what's possible with our open models. At Google DeepMind, our commitment to open science and developer choice remains unwavering. Gemma 4 is just one step in this journey, and we are excited to keep iterating alongside you. To celebrate, here is a look at the real-world, global impact Gemma is having, and how it’s empowering the next generation of builders 👇 Thank you to everyone who has downloaded, fine-tuned, and built with Gemma. We can't wait to show you what's next.

Celebrating the milestone of a massive 150+ million downloads of Gemma 4 with the release of the new Gemma 4 12B model! It's incredibly powerful for such a small model and it’s tiny enough to run locally on a laptop with just 16GB VRAM. Apache 2.0 license - happy building!

Meet Gemma 4 12B! A unified, encoder-free multimodal model designed to bring high-performance intelligence directly to your laptop, and released under an Apache 2.0 license. Bridging the gap between edge efficiency and advanced reasoning. Here is what’s new with Gemma 4 12B: 👇

introducing Managed Agents on the Gemini API - in one API call, you get agent that comes with a remote Linux environment hosted by Google, ready to scale - you can define custom instructions, skills, and tools in Markdown