Product Launch3 min read

llama.cpp and ggml reach version 1.0 with API stability guarantees

The ggml-org team has released v1.0.0 of libllama and libggml and adopted semantic versioning. Breaking changes to the public C API will now only come with a major version bump, giving apps built on local inference a stable base.

AE
Aidose Editorial
Nov 16, 2026

The maintainers of llama.cpp have released version 1.0.0 of its two core libraries, libllama (llama.cpp) and libggml (ggml), according to a November 13 post on the Hugging Face blog by Daniel Bevenius, Georgi Gerganov and merve of ggml-org.

With the release, the projects adopt semantic versioning. "Moving forward, both llama.cpp and ggml will reflect API and ABI incompatibilities with major-version bumps," the team wrote. A major version change marks a breaking change to the public C API in the ./include directory. Minor versions cover backward-compatible features, backend updates, new model support and API additions, and patch versions are bug fixes with no API change.

According to the post, llama.cpp now supports more than 150 architecture variants and all mainstream speculative decoding methods. The ggml engine runs on dedicated hardware backends, each of which can be built as a dynamic library and loaded at runtime.

The libraries ship through Docker and Debian on Linux, Homebrew, conda-forge, MacPorts, Winget, and an XCFramework for iOS and macOS, with pre-release builds for Linux, macOS, Windows, Android and iOS. The team uses Hugging Face Buckets and HF Jobs to scale its CI, and NVIDIA, AMD, Qualcomm, Arm, IBM and Intel provide hardware for self-hosted test runners. The llama.cpp team is part of Hugging Face.

For developers who embed llama.cpp in desktop apps, mobile apps or inference servers, the version numbers now say whether an upgrade can break their build. Until now, tracking the fast-moving master branch meant watching for API changes by hand. Packagers and downstream projects can now pin to a major version.

AE
Aidose Editorial
Nov 16, 2026 · 3 min read
Back to News