By developers.googleblog.com
Publication Date: 2026-04-02 00:00:00
Today Google DeepMind was launched Gemma 4a family of cutting-edge open models that redefine what’s possible on your own hardware. Gemma 4 is now available under the Apache 2.0 license, providing developers with a powerful toolkit for on-device AI development. With Gemma 4, you can now go beyond chatbots and create agents and autonomous AI use cases that run directly on the device. Gemma 4 enables multi-stage planning, autonomous action, offline code generation and even audio-visual processing, all without special fine-tuning. It is also designed for a global audience and supports over 140 languages.
Unfortunately, your browser does not support playback of this video
Gemma 4 enables visual processing and support in >140 languages
We’re excited to announce that starting today, you can experience Gemma 4’s extensive features on the Edge! Access Android’s built-in Gemma 4 model via the new AICore Developer Preview or…