llamafile
Project record
Maintained by Mozilla.ai16
llamafile packages an open model's weights and a llama.cpp-based inference runtime into a single executable that runs locally on several operating systems without installation. It combines llama.cpp with Cosmopolitan Libc and includes whisperfile, a single-file speech-to-text tool built on whisper.cpp.13
- Website: Website (external site: mozilla.ai)
- Repository: Repository (external site: github.com)
- Documentation: Documentation (external site: docs.mozilla.ai)
- License: LICENSE (external site: github.com)
- Release notes: Reviewed release (external site: github.com)
Availability and license
Overall availability
Source code, release binaries, and documentation are public. Pre-built llamafiles for example models are hosted on Hugging Face.15
Availability is separate from permission: read the license before using or redistributing.
Component reuse rights
- weights
- Unknown — no complete fact-level rights review
- code
- Reviewed qualifying license recorded — check scope and conditions
- data
- Unknown — no complete fact-level rights review
- documentation
- Unknown — no complete fact-level rights review
No complete system-rights review is recorded for this release.
Public materials checklist
| Item | Status | Notes and evidence |
|---|---|---|
| Source codeIs the source code publicly readable? | Public | Public repository in the mozilla-ai GitHub organization.1 |
| DocumentationIs user documentation published? | Public | Official documentation covers quick start, pre-built llamafiles, creating llamafiles, source installation, supported systems, and whisperfile.31 |
| InstallationAre installation instructions or packages publicly available? | Public | The quick start downloads a pre-built llamafile, marks it executable, and runs it; on Windows the file is renamed with an .exe extension. Executables over 4 GB cannot run on Windows, so larger models use the llamafile binary with external GGUF weights. Source installation is documented.1 |
| Supported platformsAre supported operating systems or hardware documented? | Public | Documented for Linux, macOS (GPU support on ARM64 only), Windows 10 and later (AMD64), FreeBSD, NetBSD (AMD64), and OpenBSD 7.0 to 7.4 (AMD64), on AMD64 CPUs with AVX and ARMv8 CPUs. GPU backends include Apple Metal, NVIDIA CUDA, AMD ROCm/HIP, and Vulkan; the docs say the 0.10 series has not yet been tested on every GPU and platform.4 |
| Release statusAre versioned releases published? | Public | Reviewed release llamafile v0.10.6, published September 15, 2026. Versions from 0.10.0 use a new build system aligned with recent llama.cpp; earlier versions remain on the releases page. This entry describes the project rather than one release.51 |
What it is useful for
Distributing and running open language models on a local machine as one self-contained file, or running the llamafile binary with separately downloaded GGUF weights.1
Run and use notes
Organization context
Provenance and derivatives
llamafile combines llama.cpp with Cosmopolitan Libc, and whisperfile builds on whisper.cpp. The README badges pin specific llama.cpp and whisper.cpp commits.1
- Derived from: llama.cpp — Inference engine combined into llamafile.
- Derived from: whisper.cpp (external site: github.com) — Basis of whisperfile.
- Derived from: Cosmopolitan Libc (external site: github.com) — Used to build single executables that run on several operating systems.
U.S. eligibility
Eligible · basis: U.S.-governed project
Mozilla.ai announced in October 2025 that the llamafile codebase had joined its GitHub organization, and the README says the former Mozilla Builders project is now revamped by Mozilla.ai. The LICENSE names the Mozilla Foundation as copyright holder. Mozilla.ai (MZL.AI, PBC) gives a San Francisco address, and the Mozilla Foundation's audited financial statements list it as a wholly owned subsidiary of the foundation, a California not-for-profit corporation headquartered in San Francisco (see the Mozilla.ai record).61278
Sources
This listing is not an endorsement, a safety assessment, or a federal approval.