Local LLM runtime
llama.cpp
llama.cpp is a MIT-licensed local-LLM server-capable runtime for Linux · macOS · Windows, running on CUDA · Metal · CPU · Vulkan · ROCm, with an OpenAI-compatible API. Ease of use 3/5 in our sourced ratings.
- License
- MIT
- Operating systems
- Linux · macOS · Windows
- Backends
- CUDA · Metal · CPU · Vulkan · ROCm
- Install
- source / binary
- Ease of use
- 3/5
- Latest version
- b9935
- GitHub stars
- 119,748
- Engine
- ggml
- Graphical app (GUI)
- Server mode
- OpenAI-compatible API
- CPU offload
- Multi-GPU
- Speculative decoding
- LoRA support
- KV-cache quantization
As of 2026-07-11 · source · GitHub
Quant-format support
What the feasibility gate checks before any VRAM math — per format and backend.
| Format | Backend | Supported |
|---|---|---|
| gguf | CPU | yes |
| gguf | CUDA | yes |
| gguf | Metal | yes |
| gguf | ROCm | yes |
| gguf | Vulkan | yes |
| mlx | Metal | no |
Software that runs on llama.cpp
Frontends, assistants and harnesses that can use llama.cpp as their local backend.
Recommended in workflows
Will your machine run a model with llama.cpp?
Check with llama.cpp preselected →Support facts come from each runtime's own documentation, stamped and dated — how the math works.