Google Releases Gemma 4 Under Apache 2.0 License

Google has launched its Gemma 4 family of open-weight AI models, now released under the permissive Apache 2.0 license, removing previous usage restrictions that hindered enterprise adoption. This move positions Gemma 4 to compete directly with other leading open-weight models like Mistral and Qwen, eliminating legal friction for businesses. The new family includes four models across 'workstation' and 'edge' tiers, featuring advanced capabilities such as native multimodality (text, vision, audio), large context windows up to 256K tokens, and efficient Mixture-of-Experts (MoE) architecture for reduced inference costs. Google's decision to adopt a standard open-source license, especially as some competitors move towards more restrictive terms, signals a strong commitment to the open AI ecosystem.