Can I run Llama 3.2 Vision 11B on a Mac?

Llama 3.2 Vision 11B needs about 9.3 GB of memory. It fits on 15 of the 15 machines we track. The cheapest current machine that runs it at a comfortable 10+ tokens/sec is the Mac mini M6 (16GB) ($899).

How much memory does Llama 3.2 Vision 11B need?

About 9.3 GB: 7.8 GB for the 4-bit model plus room for context.

Can I run Llama 3.2 Vision 11B locally? Speed on every Mac & mini PC

Machine	Memory	Fits?	Tokens/sec	From
Mac Studio M5 Ultra (96GB)	96 GB	Yes	120+	$5,499	Check price →
Mac Studio M5 Ultra (256GB)	256 GB	Yes	120+	$9,499	Check price →
Mac Studio M5 Max (40-core GPU, 64GB)	64 GB	Yes	63	$3,799	Check price →
Mac mini M5 Pro (24GB)	24 GB	Yes	32	$1,699	Check price →
Mac mini M5 Pro (64GB)	64 GB	Yes	32		Check price →
Mac mini M4 Pro (64GB) older	64 GB	Yes	28		Check price →
NVIDIA DGX Spark (128GB)	128 GB	Yes	25		Check price →
Ryzen AI Max+ 395 mini PC (128GB)	128 GB	Yes	23		Check price →
Mac mini M6 (16GB)	16 GB	Yes	17	$899	Check price →
Mac mini M6 (32GB)	32 GB	Yes	17		Check price →
Mac mini M4 (16GB) older	16 GB	Yes	12		Check price →
Mac mini M2 (16GB) older	16 GB	Yes	10		Check price →
Beelink SER7 (Ryzen 7 7840HS, 32GB)	32 GB	Yes	6.9		Check price →
Minisforum UM790 Pro (Ryzen 9 7940HS, 32GB)	32 GB	Yes	6.9		Check price →
GEEKOM A6 (Ryzen 7 6800H, 32GB)	32 GB	Yes	5.9		Check price →

Around 10 tokens/sec reads comfortably; under 5 feels slow. How we estimate.

Other llama3.2-vision sizes

Llama 3.2 Vision 90B

Embed this table

Free for blogs, READMEs and model cards. Updates automatically.

<iframe src="https://vsmacs.com/embed/llama3-2-vision-11b" width="100%" height="420" style="border:0" loading="lazy" title="Llama 3.2 Vision 11B speed by machine"></iframe>