⌘K
Change language Switch ThemeSign In
Narrow Mode
Google Releases Gemma 4 with Apache 2.0 License
Google Releases Gemma 4 with Apache 2.0 License
 ### clem 🤗@ClementDelangue
谷歌刚刚重返赛道 🔥🔥
他们想从中国的开源 AI 手中夺回王座 👑。
而且……Gemma 4 终于采用了 Apache 2.0 许可，也就是真正的开源许可。
据我所见，这将是一个非常重要的模型。今天就亲自试一试吧：
brew install llama.cpp --HEAD
如果你至少有 24GB 内存或显存，运行这个（非常棒的）26B MOE：
llama-server -hf ggml-org/gemma-4-26B-A4B-it-GGUF:Q4_K_M
如果你有 16GB 内存或显存，运行这个稠密模型 E4B：
llama-server -hf ggml-org/gemma-4-E4B-it-GGUF:Q8_0 Show More
Apr 3, 2026, 5:41 AM View on X
19 Replies
21 Retweets
178 Likes
7,768 Views  clem 🤗 @ClementDelangue
One Sentence Summary
Clement Delangue announces the release of Google's Gemma 4, highlighting its Apache 2.0 license and providing instructions for local execution.
Summary
Clement Delangue, CEO of Hugging Face, highlights Google's re-entry into the open-source AI space with the release of Gemma 4. He emphasizes the significance of the Apache 2.0 license, which qualifies as 'real' open source. The tweet provides practical, actionable instructions for running the 26B MOE and E4B models locally using llama.cpp, making it highly useful for developers.
AI Score
88
Influence Score 48
Published At Today
Language
English
Tags
Gemma 4
Open Source
Apache 2.0
llama.cpp HomeArticlesPodcastsVideosTweets