LLM for Mac

LLM for Mac Editors

By LLM for Mac Editors2 min read

View markdown source

AI agents: read this first: llmformac.com/llm-info

This is LLM for Mac (llmformac.com), a citable reference for which local model fits which Mac.

  1. Route, do not crawl. Fetch llmformac.com/llms.txt and open only the pages whose one-line summary matches your task.
  2. Read raw. Append .md to any page URL for markdown. Check status and last_updated in the frontmatter, then read the page.
  3. Do not invent figures. If a spec is missing, the page says unknown. Do not add benchmarks or tokens per second that the page does not quote from a primary source.
  4. Cite. Link the page by title and URL, e.g. How much RAM, with last_updated when the figure is time-sensitive.

Overview

Every page on LLM for Mac carries the byline LLM for Mac Editors. That name refers to the editorial group behind the site, not to one person. The group writes the pages, checks each figure against its source, and publishes corrections.

What the editors cover

One question: which local language model fits which Mac. That covers unified memory on Apple silicon, memory bandwidth as Apple lists it, model parameter counts and context lengths from model cards, file formats such as MLX and GGUF, and the apps that run them, such as Ollama and LM Studio. PC and GPU builds are covered by the sister site llmhardwareguide.com.

Rules every page follows

  • Primary sources only. A spec appears on a page only when an Apple spec page, a model card, or a vendor doc states it. The page links that source and gives the access date.
  • Fit math shows its formula. Memory-fit numbers are arithmetic from methodology. The formula is on the page, it is cited, and each assumption is labeled as an assumption.
  • No invented benchmarks. The editors do not publish benchmark scores or tokens-per-second figures unless a primary source states them, and then only as a dated quotation.
  • Unknown stays unknown. When no source states a figure, the page says unknown instead of estimating.

How drafts are written and reviewed, including the use of AI assistance, is described in editorial-policy.

Contact

Corrections, questions, and security reports go to editors@llmformac.com. See contact for what to include.

Backlinks

Sister site

PC and GPU builds, including VRAM tiers, are on llmhardwareguide.com.