Google has introduced Gemma 4, its latest open-source AI model licensed under Apache 2.0, allowing developers complete control over their applications and data. Unlike its proprietary counterpart, Gemini, which is integrated into Google’s products, Gemma can be run locally on billions of devices, including Android smartphones and some laptop GPUs, offering a secure and private experience without the need for an internet connection.
Gemma 4 boasts advanced reasoning capabilities, improvements in math and instruction-following tasks, and support for agentic workflows. It can also process audio and video for speech recognition and visual interpretation, and is available in various sizes ranging from two to 31 billion weights. With support for over 140 languages and a substantial context window, Gemma 4 enables developers to integrate AI into their applications without costly subscriptions. Discover Gemma 4 today on platforms like Google AI Studio, Hugging Face, and Kaggle.
Source link
