llama.cpp and ggml reach version 1.0.0 for local AI models
The open-source engine behind many locally run AI models hits 1.0, supporting more than 150 model architecture variants across major hardware.
By KreatorFlow News
1 min1 source
The ggml-org team announced version 1.0.0 of llama.cpp and its underlying tensor library ggml on November 13, 2026, in a post on the Hugging Face blog.
What happened
llama.cpp is an engine for running large language models locally, written in C/C++. ggml is the machine learning tensor library it is built on. Both are now at v1.0.0.
The team says llama.cpp supports more than 150 model architecture variants, along with mainstream speculative decoding methods. Hardware support comes from NVIDIA, AMD, Qualcomm, Arm, IBM and Intel, and the project runs its tests on self-hosted machines across those platforms, using Hugging Face Buckets and Jobs for capacity.
Builds are distributed through Docker, Debian, Homebrew, conda-forge, MacPorts, Winget and an XCFramework for Apple platforms.
Why it matters for creators
Many local AI tools, from offline chat assistants to prompt helpers, run on llama.cpp. A 1.0 release signals a stable base for the apps built on top of it, which matters for creators who want to run models on their own computer for privacy or cost.
Sources
What this story is based on. Links open the original pages.
