Who's who · Open-source projects
llama.cpp
Open-source C/C++ engine for running language models locally on everyday hardware
At a glance
- Founded
- March 2023
- People
- Georgi Gerganov, Creator and lead maintainer; founder of ggml.ai
- Known for
- llama.cpp, ggml, llama-server
- Licence
- MIT
- Funding
- Maintained by ggml.ai, part of Hugging Face since Feb 2026
- GitHub stars
- 130,000+ (Oct 2026)
- Parent
- Hugging Face (since Feb 2026)
- Website
- https://github.com/ggml-org/llama.cpp
llama.cpp is an open-source engine for running large language models on ordinary computers, written in plain C and C++ with no outside dependencies. It is built on ggml, a tensor library from the same team that also underpins whisper.cpp. Quantisation, which stores model weights at lower precision, from 1.5-bit to 8-bit lets large models run in far less memory. It supports a wide range of hardware, including Apple silicon through Metal, NVIDIA GPUs through CUDA, AMD GPUs, Vulkan, Intel GPUs and ordinary CPUs. It ships a command-line tool and an OpenAI-compatible server with a built-in web interface, and can download models directly from Hugging Face.
Georgi Gerganov started the project in March 2023; it now has more than 130,000 GitHub stars. Gerganov founded ggml.ai in 2023 to support the work, with pre-seed funding from Nat Friedman and Daniel Gross.
In February 2026 ggml.ai joined Hugging Face. The team works full time on ggml and llama.cpp, the community keeps control of technical decisions, and the code stays under the MIT licence. NVIDIA has agreed to buy Hugging Face, a deal expected to close in the first half of 2027.
Sources
- ggml-org on GitHub, ggml.ai joins Hugging Face to ensure the long-term progress of Local AI, 2026-02-20
- Hugging Face, GGML and llama.cpp join HF to ensure the long-term progress of Local AI, 2026-02-20
- ggml.ai, ggml.ai, 2026-10-02
- GitHub, ggml-org/llama.cpp repository, 2026-10-02
Facts last checked 2026-10-02