Google Unveils Gemma 4 for Advanced AI Workflows

Google launches Gemma 4, a new open AI model family built for advanced reasoning, coding and multimodal workflows.

Google has launched Gemma 4, its latest family of open AI models designed for advanced reasoning, coding and multimodal workflows, marking a major expansion of its developer-focused artificial intelligence ecosystem.

According to Google DeepMind, Gemma 4 is purpose-built for agentic workflows and advanced reasoning tasks, with four model sizes tailored for different use cases, from smartphones and laptops to high-end workstations and data centres.

The model lineup includes Effective 2B (E2B), Effective 4B (E4B), a 26B Mixture of Experts model, and a 31B dense model. Google said the larger models are capable of handling sophisticated coding and reasoning tasks, while the smaller variants are optimised for memory and compute efficiency on edge devices.

Gemma 4 also supports multimodal processing, enabling developers to work with text, images and video inputs. The company said the models are released under the Apache 2.0 license, making them more accessible for commercial and research use.

Google noted that the Gemma ecosystem has surpassed 400 million downloads since its initial launch, with over 100,000 community-built variants now in circulation.