Skip to content

Reading the repo

Opening the box…

How to install llama.cpp on Linux (ggml-org/llama.cpp) · gitmanual

How to install on Linux

ggml-org/llama.cpp

LLM inference in C/C++

PythonC/C++ (CMake)#ggml

Difficulty: Moderate
Time: about 15 min
Stars: ★ 133,555
Activity: Updated this month
License: MIT
Open issues: 2,589
Got it running: Be the first
Looked up: 6 h ago

Saved setup from 6 h ago

Start here

Website ↗llama.appThe project's own site.

Setup script in the repo

The maintainers included scripts/install-oneapi.bat. Run it from the cloned folder after reading it.

.\scripts\install-oneapi.bat

Will it work on my device?

Yes, it installs on your Linux

A ready-made Linux download is in the latest release.

Checking for

macOS Windows Linux

Download for Linux

b12173 (pre-release) · 11/6/2026 ↗

cudart-llama-b12173-bin-ubuntu-cuda-12.8-x64.tar.gz

tar.gz · x64 · 594.4 MB · best match

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/cudart-llama-b12173-bin-ubuntu-cuda-12.8-x64.tar.gz"
tar -xzf "cudart-llama-b12173-bin-ubuntu-cuda-12.8-x64.tar.gz"

Other Linux downloads (17)

cudart-llama-b12173-bin-ubuntu-cuda-13.5-x64.tar.gz

tar.gz · x64 · 444.1 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/cudart-llama-b12173-bin-ubuntu-cuda-13.5-x64.tar.gz"
tar -xzf "cudart-llama-b12173-bin-ubuntu-cuda-13.5-x64.tar.gz"

llama-b12173-bin-ubuntu-cuda-12.8-x64.tar.gz

tar.gz · x64 · 188.2 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-cuda-12.8-x64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-cuda-12.8-x64.tar.gz"

llama-b12173-bin-ubuntu-cuda-13.5-x64.tar.gz

tar.gz · x64 · 169.7 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-cuda-13.5-x64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-cuda-13.5-x64.tar.gz"

llama-b12173-bin-ubuntu-openvino-2026.4.1-x64.tar.gz

tar.gz · x64 · 122.6 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-openvino-2026.4.1-x64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-openvino-2026.4.1-x64.tar.gz"

llama-b12173-bin-ubuntu-rocm-10.0-x64.tar.gz

tar.gz · x64 · 255.5 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-rocm-10.0-x64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-rocm-10.0-x64.tar.gz"

llama-b12173-bin-ubuntu-sycl-fp16-x64.tar.gz

tar.gz · x64 · 67.2 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-sycl-fp16-x64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-sycl-fp16-x64.tar.gz"

llama-b12173-bin-ubuntu-sycl-fp32-x64.tar.gz

tar.gz · x64 · 67.1 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-sycl-fp32-x64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-sycl-fp32-x64.tar.gz"

llama-b12173-bin-ubuntu-vulkan-x64.tar.gz

tar.gz · x64 · 35.9 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-vulkan-x64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-vulkan-x64.tar.gz"

llama-b12173-bin-ubuntu-x64.tar.gz

tar.gz · x64 · 23.0 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-x64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-x64.tar.gz"

llama-b12173-bin-ubuntu-26.04-s390x.tar.gz

tar.gz · 15.7 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-26.04-s390x.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-26.04-s390x.tar.gz"

llama-b12173-bin-ubuntu-s390x.tar.gz

tar.gz · 17.9 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-s390x.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-s390x.tar.gz"

llama-b12173-bin-ubuntu-zdnn-s390x.tar.gz

tar.gz · 13.8 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-zdnn-s390x.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-zdnn-s390x.tar.gz"

cudart-llama-b12173-bin-ubuntu-cuda-13.5-arm64.tar.gz

tar.gz · ARM64 · 559.9 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/cudart-llama-b12173-bin-ubuntu-cuda-13.5-arm64.tar.gz"
tar -xzf "cudart-llama-b12173-bin-ubuntu-cuda-13.5-arm64.tar.gz"

llama-b12173-bin-linux-arm64-snapdragon.tar.gz

tar.gz · ARM64 · 25.6 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-linux-arm64-snapdragon.tar.gz"
tar -xzf "llama-b12173-bin-linux-arm64-snapdragon.tar.gz"

llama-b12173-bin-ubuntu-arm64.tar.gz

tar.gz · ARM64 · 18.7 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-arm64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-arm64.tar.gz"

llama-b12173-bin-ubuntu-cuda-13.5-arm64.tar.gz

tar.gz · ARM64 · 164.9 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-cuda-13.5-arm64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-cuda-13.5-arm64.tar.gz"

llama-b12173-bin-ubuntu-vulkan-arm64.tar.gz

tar.gz · ARM64 · 29.2 MB

Move the extracted binary onto your PATH (e.g. /usr/local/bin).

curl -LO "https://github.com/ggml-org/llama.cpp/releases/download/b12173/llama-b12173-bin-ubuntu-vulkan-arm64.tar.gz"
tar -xzf "llama-b12173-bin-ubuntu-vulkan-arm64.tar.gz"

Did you get it running?

No reports for Linux yet.

Still stuck? Ask an AI

Opens a chat with this manual already filled in, for Linux. It uses your own account; gitmanual sends nothing.

Ask Claude Ask ChatGPT

Parts list

1×Python >=3.10,<3.15runtime
1×C/C++ compiler runtime
1×CMake 3.14...3.28runtime
1×poetry package manager

3× settings in .env

VITE_PUBLIC_APP_NAME
VITE_DEBUG
VITE_PUBLIC_SERVER_ORIGIN

Share this manual

Is this your repo?

Add a badge to your README so newcomers find these steps.

[![Setup manual](https://gitmanual.com/ggml-org/llama.cpp/badge.svg)](https://gitmanual.com/ggml-org/llama.cpp)

Good to know

Native projects often need extra system libraries. If the build stops with a missing package, check the maintainers' build guide (see Start here).
The README mentions NVIDIA CUDA. Check its hardware section: some features (or all of them) may need an NVIDIA GPU.
Has AGENTS.md — instructions for coding agents. You can open Claude Code (or another agent) in the cloned folder and ask it to do the setup for you.
Nested projects: app/CMakeLists.txt, common/CMakeLists.txt, examples/CMakeLists.txt, examples/batched/CMakeLists.txt and 11 more

Files we read (13)

.github/workflows/behavior-check.yml

.github/workflows/build-3rd-party.yml

.github/workflows/build-and-test-snapdragon.yml

.github/workflows/build-android.yml

.github/workflows/build-apple-daily.yml

.github/workflows/build-apple.yml

CMakeLists.txt

Makefile

README.md

docs/install.md

pyproject.toml

requirements.txt

tools/ui/.env.example

Branch master · 4,074 files in the repo