ToolsNews
llama.cpp 1.0: stable API and semantic versioning
After three and a half years, llama.cpp and ggml hit version 1.0.0. The point is not a new feature but a promise: from now on, every API or ABI break comes with a major version bump, so downstream apps can upgrade predictably.
· 2 min read

What changed
The ggml-org team, now part of Hugging Face, released libllama and libggml 1.0.0 (Hugging Face). Both projects now follow semantic versioning:
- MAJOR: breaking change to the public C API in the include folder.
- MINOR: backward-compatible feature, backend update, model support or API addition.
- PATCH: bug fix with no API change.
Where llama.cpp stands
- It supports more than 150 architecture variants and all mainstream speculative decoding methods.
- ggml backends can be built as dynamic libraries and loaded at runtime.
- Official channels: pre-release builds for Linux, macOS, Windows, Android and iOS, plus Docker, Debian, Homebrew, conda-forge, MacPorts, Winget and XCFramework.
- CI scales on HF Buckets and HF Jobs, with self-hosted runners from NVIDIA, AMD, Qualcomm, Arm, IBM and Intel.
The team also notes it uses LLMs in development, while keeping control of the architecture and understanding of the code as top priorities.
Why it matters
Anyone shipping on llama.cpp used to risk silent breakage on every update. After 1.0 the version number tells you when code changes are needed, so the dependency is easier to pin and upgrade. Local, private inference is becoming a baseline requirement, and it needed a stable foundation.
Produced with the help of AI, reviewed by an editor. How we make our articles
Reviewed and edited by the Neural Daily desk. About


