Google’s AI Edge Team. Today, Google DeepMind released Gemma 24, a new family of cutting-edge open models that push the boundaries of what’s achievable on local hardware. Gemma 4 is now available under the Apache 2.0 license, providing developers with a powerful toolkit for building on-device AI applications. With Gemma 4, you can now move beyond simple chatbots to create agents and fully autonomous AI applications that run directly on-device. Gemma 4 supports multi-step planning, autonomous behavior, offline code generation, and audio-visual processing—all without requiring any specialized fine-tuning. It’s also designed for a global audience, with support for more than 140 languages. Sorry, your browser doesn’t support video playback. Gemma 4 supports visual processing and more than 140 languages. We’re thrilled to share that Gemma 43’s powerful features are now available on the edge, starting today!
Google Developers Blog