Unified Memory
Unified memory is one pool of RAM shared by the processor and the graphics cores — the reason Apple Silicon Macs can run models that would need an enormous graphics card elsewhere.
A desktop PC keeps two separate memories: the graphics card's fast VRAM and the system's ordinary RAM. Apple Silicon merges them — every chip has a single memory pool, and the GPU can address most of it. A 64 GB Mac therefore competes, for model-fitting purposes, with graphics cards far above consumer budgets.
There are two honest caveats. First, macOS caps how much of that pool the GPU may claim — roughly two-thirds to three-quarters, depending on machine size — so a 64 GB Mac does not offer 64 GB to a model. Second, memory bandwidth varies a lot across Mac tiers and generally sits below flagship graphics cards, which shows up as generation speed, not capability.
For running large models occasionally and quietly, few consumer setups match it. For maximum tokens per second, dedicated GPUs still lead.
Where you see it on the radar
Mac profiles on the radar plan with 70% of unified memory and say so on the page — the macOS cap is real, and ignoring it would oversell what fits.
Compare Macs and GPUs honestly →Related terms