Can I Run Qwen 2.5 VL on Mac Studio (M3 Ultra, 256 GB)?

Superseded model. Qwen 2.5 VL has been superseded by Qwen 3. This page is kept for reference; the newer family is a better starting point. View Qwen 3 →

Written by Jakub Rusinowski · Last updated January 25, 2025

Yes — comfortably

Yes, comfortably — Qwen 2.5 VL 72B Instruct at Q8_0 needs about 81.5 GB of the 192 GB usable on Mac Studio (M3 Ultra, 256 GB), leaving ~110.5 GB spare and running at ~5.6 tok/s (estimated), with room for about 65,536 tokens of context.

Confidence: high · Recommended quantization: Q8_0 · Estimated speed: ~5.6 tok/s

Mac Studio (M3 Ultra, 256 GB) — what it gives a model

Usable memory for models	192 GB
Memory bandwidth	819 GB/s
Form factor	Desktop
Operating system	macOS
Memory upgradeable	No — soldered

Qwen 2.5 VL on Mac Studio (M3 Ultra, 256 GB): memory by quantization

Quant	Memory needed	Fits 192 GB?	Max context	Est. speed	Download
F16	150.3 GB	✓ Yes	64K	~3 tok/s	146.8 GB
Q8_0	81.5 GB	✓ Yes	64K	~5.6 tok/s	78 GB
Q6_K	63.7 GB	✓ Yes	64K	~7.1 tok/s	60.2 GB
Q5_K_M	55.5 GB	✓ Yes	64K	~8.2 tok/s	52 GB
Q4_K_M	47.8 GB	✓ Yes	64K	~9.6 tok/s	44.3 GB
Q3_K_M	34.8 GB	✓ Yes	64K	~13.2 tok/s	31.3 GB
Q2_K	27.6 GB	✓ Yes	64K	~16.7 tok/s	24.1 GB

Which Qwen 2.5 VL sizes fit

Variant	Needs at Q4_K_M	Fits?	Est. speed
Qwen 2.5 VL 72B Instruct	47.8 GB	✓ Fits	~9.6 tok/s
Qwen 2.5 VL 7B Instruct	6.3 GB	✓ Fits	~68.2 tok/s

What to watch out for

Memory on this machine is not upgradeable, so the configuration you buy is the ceiling for every model you will ever run on it.

Mac Studio M3 Ultra 256 GB limitations

The largest single-box unified memory pool available, but at roughly half the memory bandwidth of a high-end discrete GPU — very large models load, then run slowly.
Memory is soldered; the configuration chosen at purchase is permanent.

Recommended setup

Ollama or LM Studio (Metal) — MLX for the fastest Apple-native throughput

How these numbers are calculated

Memory need = quantized weights + KV cache at 8,192 tokens (f16) + 0.8 GB runtime overhead.
256 GB unified memory at 819 GB/s, shared between CPU and GPU.
macOS reserves a share of unified memory for the system, so not all of it is available to a model.
Throughput is a memory-bandwidth roofline estimate, not a measurement. It is labelled "estimated" everywhere it appears.
KV cache is computed from this model's published attention configuration.

FAQ