Google Releases Gemma 4 Open-Weight AI Models Under Apache 2.0 License

Google released Gemma 4, a family of four open-weight AI models (2B, 4B, 26B, 31B parameters) built on Gemini 3 research, now under an Apache 2.0 license instead of Google's proprietary Gemma license, offering greater developer flexibility.

Evidence Strength

Evidence

75%Strong

Based on original reporting

Corroborated by 4 publishers

Key Development

High-significance development (rated 8/10)

Widely covered — 5 independent sources

Confirmed — verified event

Insights

First tracked

April 2, 2026

Last updated

April 14, 2026

Sources

5 sources

Related Developments

NVIDIA and Google Collaborate to Optimize Gemma 4 Models for NVIDIA GPUs 24/7 Wall St: DeepMind Acquisition Origin Story Illustrates Alphabet's Foundational AI Advantage Google Leaders Including Demis Hassabis Push Back on Claims of Uneven Internal AI Adoption Google Developing Screen-Less Fitbit Fitness Band to Compete with Whoop Gemini App Surpasses 750 Million Monthly Active Users

Sources (5)

Source Timeline

Google's Gemma Already Acts Like Gemini—Someone Made It Think Like Claude Opus Too
Decrypt·Apr 14
Google battles Chinese open-weights models with Gemma 4
Theregister·Apr 2
Google Jumps Back Into the Open Source AI Race With Gemma 4
Decrypt·Apr 2
Google releases Gemma 4 under Apache 2.0 — and that license change may matter more than benchmarks
Venturebeat·Apr 2
Google releases Gemma 4, a family of open models built off of Gemini 3
Engadget·Apr 2

View all Alphabet/Google developments