Google Launches Gemma 4, a New Open-Source Model: How to Try It

Google has unveiled its latest AI model, Gemma 4, which is now both open-weight and open-source. Released under the Apache 2.0 license, Gemma 4 is a fully open-source model, setting it apart from other frontier models. This allows users to run it locally on billions of Android devices and some laptop GPUs, granting developers complete control over their data and infrastructure. Unlike proprietary models, Gemma 4’s code and training data are shared, ensuring privacy since no data is shared with third parties. Developers can use Gemma 4 without any subscription costs, integrating AI into their applications seamlessly. Gemma 4 is designed for advanced reasoning and logic, supporting audio and visual processing, and is available in sizes up to 31 billion weights. With a context window of up to 256,000 tokens in its largest variant, Gemma 4 offers flexibility for various use cases. Open and open-source, Gemma 4 can be downloaded and modified freely, allowing personal and commercial applications without royalty requirements. It’s available for download at Google AI Studio and other platforms like Hugging Face.