Google releases Gemma 4 under Apache 2.0 license, abandoning custom terms for OSI-approved license

Google released Gemma 4 on April 2, 2026 in four sizes: E2B, E4B, 31B, and 26B (A4B MoE). This is the first Gemma release under the OSI-approved Apache 2.0 license, replacing the custom Gemma Terms of Use that had drawn criticism from the developer community. Gemma 4 supports text, audio, and image input with up to 256K context window. The community has downloaded Gemma models over 400 million times with 100,000+ variants. E2B and E4B models are optimized for edge/phone deployment and will form the basis for Gemini Nano 4 on upcoming Pixel devices. Weights are available on Hugging Face, Kaggle, and Ollama. The license change removes ambiguity around commercial use and positions Gemma as a truly open alternative, contrasting with the trend of Chinese AI labs releasing weights with restrictive custom licenses.