What’s new in Gemma 4
By Google for Developers
Summary
## Key takeaways - **First-ever Apache 2.0 release**: For the first time, Gemma 4 is released under a fully open-source Apache 2.0 license, breaking from prior Gemma licensing. [00:39] - **Built on Gemini 3 research**: Gemma 4 inherits the same world-class research and technology behind Gemini 3, designed to run locally on phones, laptops, and desktops. [00:27] - **26B MoE with 3.8B active params**: The 26B Mixture-of-Experts model activates only 3.8B parameters, making it exceptionally fast while the 31B Dense version optimizes for output quality. [01:27] - **Quarter-million token context**: The bigger model supports up to 256K tokens, letting developers analyze entire codebases and power multi-turn agentic workflows. [00:54] - **Mobile-native multimodal models**: Effective 2B and 4B models combine audio and vision for real-time processing on mobile and IoT, natively supporting 140+ languages. [01:40], [01:48] - **Native tool use for agents**: Gemma 4 includes native tool-use support so developers can build agents that plan and act autonomously on their behalf. [01:03]
Topics Covered
- Gemma 4 Released Under Open Source Apache 2.0 License
- New Gemma 4 Models Bring AI to Mobile and IoT Devices
Full Transcript
Hi, my name is Olivier, and I'm a group product manager on the Gemma team.
Since we launched our first models, the developer community has absolutely blown us away.
Over 400 million downloads, over 100,000 variants.
You've built a vibrant ecosystem around Gemma and we could not be more grateful.
We've listened very closely to what you wanted next, and today we are thrilled to announce Gemma 4.
Built from the same world-class research and technology behind Gemini 3, Gemma 4 is our family of open models, designed to run directly on the hardware you own, phone, laptops and desktop.
For the first time ever, we are releasing Gemma under an open source Apache 2.0 license.
Gemma 4 is built for the agentic era.
It can handle complex logic, multi-step planning, and agentic workflows, making optimal use of tokens for its intelligence.
The bigger model performs well with a context window of up to a quarter million tokens, allowing you to analyze entire code bases, and/or multi-turn agentic use cases.
It features native support for tool use, allowing you to build agents that plan and act on your behalf.
Let's break down the model family now.
First, a 26B Mixture of experts, and a 31B Dense models.
These provide frontier intelligence directly on your personal computer.
You can run state-of-the-art local reasoning and coding pipeline without needing to upload data outside of your control environment.
The 26B MOE with 3.8B in activated parameters is exceptionally fast, While 31B is optimized for output quality.
Then, we have our effective 2B and effective 4B models.
Engineered for maximum memory efficiency, these models bring a whole new level of intelligence to mobile and IOT devices.
With combined audio and vision support for real-time processing, we can see and hear the world.
All of this while natively supporting over 140 languages.
Now, let's test our effective 2B on multilingual and agentic task.
Hé Gemma, est-ce que tu peux trouver un restaurant français à San Francisco?
Please reply to me in English.
Amazing. We have a winner.
As open models become more central to enterprise infrastructure, security is paramount. Developed by Google DeepMind, Gemma 4 undergoes the same rigorous security protocols as our proprietary models, giving enterprise and developers a trusted foundation to build on top of.
We want you to be able to use Gemma 4 with the tools you already know and love.
You can download the weights and start experimenting today.
We cannot wait to see what you create next.
Loading video analysis...