GGUF file format specification (ggml)
Source: https://github.com/ggml-org/ggml/blob/master/docs/gguf.md
Structured fields
{
"archived": false,
"checked_at": "2026-10-17T00:00:00.000Z",
"usefulness": "Normative description of the GGUF container used by llama.cpp, stable-diffusion.cpp and ComfyUI-GGUF style loaders: single-file, mmap-able, extensible key-value metadata, and the list of quantization type names (Q4_K, IQ4_XS, MXFP4 and others).",
"source_type": "documentation",
"availability": "available",
"source_owner": "ggml-org",
"next_check_at": "2026-10-18T00:00:00.000Z",
"upstream_updated_at": "2026-10-15T13:14:12.000Z"
}Evidence claims
- [sourced_fact] Spec states GGUF is a binary format for GGML-based executors designed for single-file deployment, extensibility and mmap compatibility, succeeding GGML, GGMF and GGJT; header must be version 3. source
- [sourced_fact] Spec lists quantization types including legacy block quants (Q4_0..Q8_1), K-quants (Q2_K..Q8_K), IQ types, TQ1_0/TQ2_0 and MXFP4, and marks some older variants removed. source
- [upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.
Rights
| component | license | hosting | redistribution | modification | training | Evidence and notes |
|---|---|---|---|---|---|---|
| documentation | unknown | unknown | unknown | unknown | unknown | License of the ggml repository docs was not opened in this check; treat as unknown. |