Give Local AI More of Your Mac's Memory: The Hidden 75% GPU Cap, and How to Lift It
macOS lets local AI models use only about three quarters of an Apple silicon Mac's memory. One command (or a free menu bar app) raises the cap. What it does, how to set it with Claude, and the catch
On Apple silicon Macs the CPU and GPU share one pool of memory. But by default macOS only lets the GPU, which is what local AI models run on, use about 75% of it. A 64 GB Mac gives AI about 48 GB; a 128 GB Mac about 96 GB. That is why a 52 GB model can fail to load fully on a 64 GB Mac. One hidden setting raises that cap.
Here is what you get, how to set it up, and the catch.
What you get
- Bigger local models fit fully on the GPU instead of spilling onto the much slower CPU.
- Example: on a 64 GB Mac, raising the cap from about 48 GB to 56 GB, while leaving 8 GB for macOS.
- It works for any app that runs models on the Mac's GPU, such as Ollama, LM Studio and llama.cpp.
- No restart needed, and one command (or one click) puts it back to Apple's default.
Free, for your email
Get the full setup steps + the copy-paste setup prompt
The rest of this guide (6 more sections) opens right here. The setup prompt, one paste into Claude that sets it all up, comes by email only. Use your real address.
Unlocks: Step 1: pick a safe number · Step 2: set it with Claude (one command) · Or: the free menu bar app · The catch · A prompt to start · +1 more
50,000+ follow where these guides come from. No spam, reply “unsubscribe” any time.
