Google Unveils Gemma 3n AI Model for Phones, Laptops, Tablets
Google launched Gemma 3n, an open-source AI model for phones, laptops and tablets, enabling local processing without cloud dependency.
image for illustrative purpose
Google on Tuesday introduced Gemma 3n, an open-source AI model designed to operate on consumer devices such as smartphones, laptops and tablets.
Unveiled during the company’s I/O 2025 developer conference, Gemma 3n joins Google’s suite of on-device offerings alongside Gemini Nano. While prior versions of Google’s Gemini models have relied on cloud processing for complex tasks, Gemma 3n enables local execution by developers integrating the model into apps or operating systems.
Built in partnership with Qualcomm, MediaTek and Samsung, the model runs in preview with a memory footprint of 2 to 3 gigabytes of RAM. According to internal tests and Chatbot Arena rankings, Gemma 3n performs on par with leading AI models, including Anthropic’s Claude 3.7 Sonnet.
Gemma 3n supports multimodal inputs, processing text, audio and images instantaneously. The model can transcribe speech, translate between languages, solve mathematical queries and handle video frames in real time.
Unlike a standalone Google application, Gemma 3n is a development toolkit that software makers can deploy directly on devices. The architecture is intended to preserve privacy by keeping data on user hardware.
The preview release is available immediately for developers. Google said feedback during this phase will guide enhancements ahead of a broader rollout.
