Will it run? / Llama 3.2 Vision 90B / Mac Studio M5 Max (40-core GPU, 64GB)

Can the Mac Studio M5 Max (40-core GPU, 64GB) run Llama 3.2 Vision 90B?

Usable, but slow Yes, at about 5.4 tokens/sec (estimated). It works, but you'll wait on longer answers. It only fits if you raise the GPU memory limit, and it'll be slower than usual.

5.4estimated tokens/sec
56.5 GBmemory needed
48 GBavailable to the model
614 GB/smemory bandwidth
ollama run llama3.2-vision:90b

Check the Mac Studio M5 Max (40-core GPU, 64GB) price →

How we got this number

Each generated token reads the whole model from memory once: 55 GB. The Mac Studio M5 Max (40-core GPU, 64GB) moves 614 GB/s, so the ceiling is 11 tokens/sec. Real machines reach a fraction of that ceiling; we use 80%, calibrated against a measured Mac run. Send us your real numbers and we'll replace the estimate.

Faster machines for Llama 3.2 Vision 90B

Mac Studio M5 Ultra (96GB)18 tok/s$5,499Check price →
Mac Studio M5 Ultra (256GB)18 tok/s$9,499Check price →
NVIDIA DGX Spark (128GB)3.5 tok/sCheck price →
Ryzen AI Max+ 395 mini PC (128GB)3.3 tok/sCheck price →

Other models on the Mac Studio M5 Max (40-core GPU, 64GB)

Qwen 3 30B (A3B)Fast120+ tok/s
Qwen3.5 35B (A3B)Fast120+ tok/s
Qwen3 Coder 30B (A3B)Fast120+ tok/s
Qwen3.6 35B (A3B)Fast120+ tok/s
Qwen 3-VL 30B (A3B)Fast120+ tok/s
gpt-oss 20BFast119 tok/s