Mac RAM to AI model size — the reference table

Outlier · solo-built in Grand Rapids · published 2026-06-09 Last updated 2026-09-15

Quick answer

16 GB handles Nano 4B and Lite 9B. 32 GB gets you the 27B coding models. 64 GB is where Plus 397B becomes an option.
On disk: Nano is 2.4 GB, Core 27B is 15.1 GB, Vision 3.8 is 15.5 GB. Plus 397B wants 209 GB.
Speed on my M1 Ultra: Nano hits 71.7 tok/s, Core 27B does 20.7, Plus 397B crawls at 1.59.
Quick math: usable model size is about your unified RAM minus a few GB for macOS.

Your Mac's unified memory decides which local models you can actually run. The table below maps it all out: what fits in your RAM, how big each model is on disk, plus the generation speed I measured on an M1 Ultra. Use it to size up the Mac you own, or the one you're about to buy.

RAM → what you can run

Unified RAM	Typical Mac	Models that run
16 GB	MacBook Air	Nano 4B, Lite 9B
24 GB	MacBook Pro	+ Quick 26B, Core 27B, Vision 3.8 27B (tight)
32 GB	MacBook Pro	27B coding + Vision, comfortably
64 GB	Mac Studio / MBP	All tiers, including Plus 397B
96+ GB	Mac Studio	Plus 397B with headroom for long context

Model size and speed (measured, M1 Ultra)

Receipts: Decode speeds are my own runs on a 64 GB M1 Ultra, the 2026-06-09 batch, and the table records that batch rather than being silently refreshed — where a figure has been re-measured since, the row says so and gives both. Disk sizes are read from the shipping tier catalog (checked 2026-09-15): Nano 2.37 GB · Lite 5.04 GB · Quick 15.61 GB · Core 15.13 GB · Vision 3.8 15.5 GB · Plus 209 GB. Corrected on the 2026-09-15 read: the quick answer said Plus ran at 2.1 tok/s while this table said 1.59 — the table was right, and a page that disagrees with itself is worse than one that is simply out of date.

Model	Params	Disk	Decode tok/s
Nano	4B	2.4 GB	71.7
Lite	9B	5 GB	53.4
Quick	26B MoE	15.6 GB	14.6 — 43.6 when re-measured 2026-08-24 on v1.11.804; this table records the 2026-06-09 run
Core	27B	15.1 GB	20.7
Vision	35B MoE	19 GB	16.3 — this row is the 2026-06-09 run on Vision 3.6; the shipping tier is now Vision 3.8, 27B dense, 15.5 GB, not re-measured here
Plus	397B MoE	209 GB	1.59

Numbers are from an M1 Ultra, MLX 4-bit, batch size 1. Newer chips (M2/M3/M4) run faster, but the ranking stays the same.

How a 209 GB model runs on a 64 GB Mac

Plus 397B is 209 GB on disk. That's way past 64 GB of RAM, and yet it runs. The trick is that it's a Mixture-of-Experts model, so only a slice of its parameters fire on any given token. Outlier's V9 paged engine holds the active experts in memory and streams the rest off the SSD, which keeps peak memory hovering around 11 GB. See how paged MoE inference works.

Questions about the RAM-to-model-size table

What size AI model can my Mac run?

Roughly your unified RAM minus a few GB for macOS. 16 GB runs 4B–9B models, 32 GB runs 27B-class models, and 64 GB runs a 397B Mixture-of-Experts model via paged streaming.

How fast is local AI on a Mac?

On an M1 Ultra: about 71.7 tok/s for a 4B model, 20.7 for a 27B model, and 1.59 for the 397B model via paged streaming. Newer chips are faster. Speed depends mostly on model size and chip generation.

How much disk do AI models need?

From about 2.4 GB for a 4B model to 209 GB for the 397B model. Most users keep a couple of small models plus one 15 GB coding model. You only download the tiers you use.

Try Outlier free

Free Nano + Lite — local, private, no account. Pro is a one-time $249 and adds everything (all 6 model tiers incl. Plus 397B). Founders Lifetime is $249 once. Apple Silicon only.

Download for Mac