Developers2 min
llama.cpp, a popular tool for running AI models at home, reaches version 1.0
The ggml-org team has released version 1.0.0 of llama.cpp and of ggml, the machine learning library it is built on. From now on, any change that breaks the public interface will bring a new major version number. The team says llama.cpp supports more than 150 model architecture variants. It is now distributed through Homebrew, Docker, Debian, conda-forge, Winget and other channels. Hugging Face also announced Llama desktop apps for macOS and Windows built on llama.cpp.
Do anything? If you build software on llama.cpp, version numbers now tell you when an update will break your code. Everyone else can keep using it as before.