📦 Ollama – v0.11.0
Hey AI Developers! 🤖
Ollama v0.11 is here, bringing OpenAI’s new gpt-oss models (20B & 120B) to your local machine! 🎉
These models unlock powerful capabilities like function calling, web browsing, Python tools, and structured outputs – perfect for building amazing applications. Plus, they utilize a new MXFP4 quantization format, letting you run the 20B model on just 16GB of RAM and the 120B on an 80GB GPU! 🤯
It’s fully customizable, Apache 2.0 licensed, and includes `kvcache` improvements for better performance.
Get it now: https://ollama.com/download
Run them with:
“`
olama run gpt-oss:20b
olama run gpt-oss:120b
“`
Happy building! ✨