Google Launches Gemini 4 AI Models with New License
April 2, 2026 at 16:01
0
✦ AI Summary
- Gemini 4 introduces local usage optimization for developers
- Switching to Apache 2.0 license enhances model accessibility
- New models boast reduced latency and improved performance
Google has unveiled its latest version of AI models, Gemini 4, which mark a significant enhancement over previous iterations. Recognizing developer concerns regarding licensing, Google has transitioned from its custom Gemma license to the more open Apache 2.0 license.
Gemini 4 comes with four variants, optimized for local deployment on powerful hardware like the Nvidia H100 GPU. The models are designed to deliver high performance and reduced latency, enhancing the user experience. The 26B Mixture of Experts variant employs an innovative parameter activation strategy that boosts tokens-per-second, while the 31B Dense variant focuses on quality, allowing developers to tailor it for specific applications.
Share: