Google Gemma 4 Goes Open-Source Under Apache 2.0: Run Frontier-Class AI Locally Without Internet

Google released Gemma 4, a family of open-weight AI models available for free under the Apache 2.0 license—meaning anyone can download, modify, and deploy them commercially. The models range from compact edge-device versions to data-center-scale reasoning models, all supporting advanced multimodal inputs (text, images, code) and agentic workflows. Critical for international access: Gemma 4 supports over 140 languages and runs entirely on-device without internet connectivity, removing the data sovereignty barriers that prevent cloud AI adoption in regulated industries across Europe, Asia, and the Middle East. On-device deployment also enables AI in areas with unreliable connectivity across Africa and Southeast Asia. Available immediately via Hugging Face and Google AI Studio; enterprise-grade support through Google Cloud.