Popular: Data updated October 24, 2026

Categories

Internet1,352
Communication321
Audio & video897
Graphics & design538
Office & productivity1,118
Developer tools3,128
AI tools525
Security & privacy385
System tools3,112
Games522
Education & science239
Network & IT283

Download llama.cpp b9994 for Windows

LLM inference in C/C++

llama.cpp is an open-source software library that performs inference on various large language models such as Meta's Llama model. It is co-developed alongside the Georgi Gerganov Machine Learning</span>"}]]}'>GGML project, a general-purpose tensor library.

Command-line tools are included with the library, alongside a server with a simple web interface.

llama.cpp has been considered as the de facto standard as the core of almost all local inference tools, including Ollama and LM Studio.

Towards the end of September 2022, Georgi Gerganov started work on the GGML library, a C library implementing tensor algebra. Gerganov developed the library with the intention of strict memory management and multi-threading. The creation of GGML was inspired by Fabrice Bellard's work on LibNC.

Before llama.cpp, Gerganov worked on a similar library called whisper.cpp which implemented Whisper, a speech to text model by OpenAI.

From the Wikipedia article “Llama.cpp”, CC BY-SA 4.0

What is llama.cpp?

llama.cpp is an open-source program for Windows from ggml. It is written in C++, C, and Python. Besides Windows, it is available for Linux, macOS and FreeBSD.

Installation and requirements

Official installers exist for 64-bit as Portable. It runs without installation (portable).

Updates and version history

The current version is b9994, released on July 14, 2026. 436 versions are listed since July 15, 2026, each with links to its own installer files. That is about 303 releases a year.

Popularity and file types

The source code on GitHub (ggml-org/llama.cpp) has 131,998 stars. Wikipedia has an article about it in 7 of our languages.

Project description on GitHub: “LLM inference in C/C++”

Pros / Cons

Pros

  • Open source: the code can be inspected and built by anyone
  • Portable: runs without installation
  • Frequent updates (about 303 a year)
  • Also runs on Linux, macOS and FreeBSD
  • Large user base

Cons

  • No native ARM64 build

In short: llama.cpp is an open-source Windows tool for 64-bit PCs. It is actively maintained.

Is llama.cpp safe to download?

  • The file is served by github.com, the publisher’s own domain
  • The download link uses HTTPS
  • A SHA256 hash is published, so you can verify the file: E21E93ED9E7A65682AC562CA48F97F7D0158DA4826E617613F22F417548CF59F · Open the VirusTotal report for this hash
  • The link has not been checked yet
  • The source code is public
  • ExeAtlas does not host, wrap or modify installers

Related guide: How to check a file’s SHA256 hash on Windows

Recent versions

All 436 versions ›
2025: 188 188 25 2026: 248 248 26

436 versions over 1 year · about 303 a year

VersionDateSizeFiles
b9994Latest July 14, 2026 — 64-bit ZIP
b9982 July 13, 2026 — 64-bit ZIP
b9969 July 12, 2026 — 64-bit ZIP
b9957 July 10, 2026 — 64-bit ZIP
b9947 July 10, 2026 — 64-bit ZIP
b9935 July 9, 2026 — 64-bit ZIP
b9910 July 8, 2026 — 64-bit ZIP
b9893 July 7, 2026 — 64-bit ZIP

Install from the command line

Silent install ›
PowerShell
wingetwinget install --id ggml.llamacpp -e

Related guide: How to use winget: install and update apps

Information

Publisher
ggml
Licence
MIT
Price
Free
Source code
github.com/ggml-org/llama.cpp
Written in
C++, C, and Python
Requires
Windows 11 / 10
Architecture
64-bit
Installer
Portable
Other platforms
Linux, macOS, FreeBSD
Version
b9994
Released
July 14, 2026
Added to the catalogue
May 21, 2025
Versions listed
436
Updates / year
303.3
GitHub stars
131,998
GitHub forks
24,671
Website
github.com
winget ID
ggml.llamacpp

Ratings and reviews

–
0 ratings
50
40
30
20
10
Rate this app

No reviews yet. Be the first to rate it.

Frequently asked questions

Is llama.cpp free?

Yes. llama.cpp is open-source software under the MIT licence.

Is llama.cpp safe to download?

The download comes from github.com, which belongs to the publisher; ExeAtlas does not host or change the file. Compare its SHA256 hash with the one shown here, or open the VirusTotal report for that hash.

How do I install llama.cpp with winget?

Open PowerShell or Windows Terminal and run the winget command shown on this page. winget downloads the official installer and checks its SHA256 hash before running it.

Does llama.cpp run on ARM (Windows on Arm)?

There is no ARM64 build; Windows 11 on Arm can run the 64-bit version through emulation.

Where can I download an older version of llama.cpp?

The versions page lists all 436 versions with links to their official installer files, back to July 15, 2026.

Can I use llama.cpp without administrator rights?

Yes, it is portable: unpack and run it.

This page in other languages

Report an error