Google Launches Gemma 4 For Advanced On-Device AI

(RTTNews) - Google has launched Gemma 4, which goes beyond chatbots and creates AI agents that can plan tasks, take actions on their own, generate code even without internet access, and process audio and video. It supports over 140 languages, making it useful for users worldwide.

Developers can access Gemma 4 on Android through the new AICore Developer Preview or use Google AI Edge to build AI-powered features across mobile, desktop, and edge devices.

Gemma 4 uses tools to fetch and process new information. It can turn long text or videos into short summaries or study notes, create charts from data, and work with other AI tools like text-to-speech, image generation, and music creation. Instead of switching between multiple apps, users can handle complex tasks and even build their own apps just by interacting with Gemma 4.

Gemma 4 is available in four different sizes, namely Effective 2B (E2B), Effective 4B (E4B), 26B Mixture of Experts (MoE), and 31B Dense, designed to do more than just chat—they can handle complex tasks and support advanced AI workflows.

The smaller models, E2B and E4B, are built for use directly on devices like phones and laptops. They focus on fast performance, support for different types of data, and smooth integration, instead of just having a large number of parameters.

Google is also launching a new Python package and command-line tool, making it easier for developers to test Gemma and build AI-powered applications, including for IoT devices.