Gemma 4 Drops Restrictive License for Apache 2.0

Google DeepMind has released Gemma 4 under the Apache 2.0 license, eliminating the custom licensing restrictions that had pushed many enterprises toward competitors like Mistral and Qwen. The previous Gemma license required legal review, flagged compliance concerns, and included restrictions on redistribution and commercial deployment. The shift to Apache 2.0 — the standard permissive license used across most of the open-weight AI ecosystem — removes that friction entirely. For enterprise teams that had been waiting for Google to match competitors on licensing terms, the change may prove more significant than any benchmark improvement.