THE HUB · TRACKER · PUBLIC

Models you can run yourself

A model, every quantisation somebody made of it, and what each one costs to run.

2 sources11,119 claimsjoined on hfpublished 5d ago

Its sources

Source	What it adds	Claims
Hugging Face modelszetlyn/models-hf	The model itself: its licence, its task and how many people fetch it.	1,654
GGUF quantisationszetlyn/models-gguf	Which quantisations exist, which is what decides whether it runs on the machine somebody has.	9,465

What it promises

fresh within	24h
covers	Every GGUF repository created since 2026-08-23, and the base model behind each one. 2,684 of those bases are named; without a Hugging Face token the run is throttled off after roughly 600 to 750 of them, so the model side of the join is partial and the scope says so on its front page.
excludes	Ollama, whose library names no Hugging Face repository to join on. Benchmarks, which have no machine-readable source.

What it compares

Each of these is held against every source that says it; a difference beyond what is allowed is a conflict. Everything else is shown side by side.

Property	Compared
`licence`	on the scale apache-2.0 > mit > llama3.1 > gemma > other > unknown

Versions

Version	Published	Claims
`dae8cbf7d033`	2026-09-29 · 5d ago		latest

A subscriber holds one of these and is told of the next; an update fetches only what changed between them.

Use it

on your own machine

$ zetlyn tracker subscribe zetlyn/local-models
# the statement, and every source it names

Version dae8cbf7d033ea3ffaed70b0 · signed ed25519:af9ffd7b7435cdad9724db5feea1a6f55ff38ea13b860ab62260f34f59a131ab · manifest.json

Its statement

name: zetlyn/local-models
title: Models you can run yourself
about: A model, every quantisation somebody made of it, and what each one costs to run.
sources:
- source: zetlyn/models-hf
  priority: primary
  why: "The model itself: its licence, its task and how many people fetch it."
- source: zetlyn/models-gguf
  priority: high
  why: Which quantisations exist, which is what decides whether it runs on the machine somebody has.
identified_by:
- hf
align:
  licence:
    scale:
    - apache-2.0
    - mit
    - llama3.1
    - gemma
    - other
    - unknown
view:
  columns:
  - kind
  - licence
  - downloads
  - known
  facets:
  - kind
  - source
  - licence
  - task
  sort: downloads desc
  named:
  - name: quantised-and-open
    title: Quantised, and openly licensed
    where: licence=apache-2.0
promise:
  fresh_within: 24h
  covers: Every GGUF repository created since 2026-08-23, and the base model behind each one. 2,684 of those bases are named; without a Hugging Face token the run is throttled off after roughly 600 to 750 of them, so the model side of the join is partial and the scope says so on its front page.
  excludes: Ollama, whose library names no Hugging Face repository to join on. Benchmarks, which have no machine-readable source.