Summary (tl;dr)

Google DeepMind has launched Gemma 4, its newest family of open-source AI models, offering enhanced capabilities for reasoning and multimodal understanding, and released under a permissive Apache 2.0 license.

Essential Background

Google introduced the Gemma series in February 2024 as a lightweight, open-source alternative to its proprietary Gemini models, aiming to make advanced AI more accessible to developers. Subsequent versions, Gemma 2 and Gemma 3, were released in June 2024 and March 2025, respectively. This initiative marked a significant shift in Google's strategy, moving towards providing open-weight models to foster innovation within the developer community.

The Full Story

Google DeepMind officially released Gemma 4 on April 2, 2026, positioning it as their "most capable open models to date." This latest iteration introduces substantial advancements, including expanded multimodal capabilities that allow the models to process text, images, video, and audio inputs, and generate text outputs. Gemma 4 also features enhanced reasoning for complex logic and multi-step planning, along with improved support for "agentic workflows" through native function-calling. A crucial aspect of this release is its availability under the commercially permissive Apache 2.0 license, which grants developers greater flexibility for modification and redistribution without royalty requirements. The Gemma 4 family comes in four sizes, from "Effective 2 Billion" (E2B) and "Effective 4 Billion" (E4B) models optimized for mobile and edge devices, to larger "26B Mixture of Experts (MoE)" and "31B Dense" models designed for high-end workstations and servers, capable of running across various hardware including Android phones, GPUs, and Google Cloud TPUs.

Why It Matters

The release of Gemma 4 is highly significant as it democratizes access to state-of-the-art AI capabilities, enabling developers and organizations to build sophisticated applications that can run locally on a wide array of devices, from smartphones to powerful servers, free from proprietary restrictions or recurring costs. Its enhanced reasoning, extensive multimodal support, and the commercially permissive open-source license are expected to drive innovation, allowing businesses and researchers to customize AI for specific use cases while maintaining full control over their data and infrastructure, a concept Google refers to as "digital sovereignty." This strategic move by Google intensifies competition within the rapidly evolving open-source AI landscape, providing a powerful and accessible alternative to other models and accelerating the development of advanced agentic AI systems.

Geographic Location

• Mountain View, Santa Clara County, California, United States (Google DeepMind's announcement of Gemma 4 release)