Can I Run Command R Family on Mac Studio (M4 Max, 64 GB)?

Newer alternative. Cohere has not released a newer model in this library. If you are choosing today, Mistral Small 4 is a comparable current option. View Mistral Small 4 →

Written by Jakub Rusinowski · Last updated April 4, 2024

Yes

Yes — Command R (35B) at Q6_K needs about 40.2 GB of the 48 GB usable on Mac Studio (M4 Max, 64 GB) (~7.8 GB spare), at ~8.6 tok/s (estimated), with room for about 8,192 tokens of context.

Confidence: high · Recommended quantization: Q6_K · Estimated speed: ~8.6 tok/s

Mac Studio (M4 Max, 64 GB) — what it gives a model

Usable memory for models	48 GB
Memory bandwidth	546 GB/s
Form factor	Desktop
Operating system	macOS
Memory upgradeable	No — soldered

Command R Family on Mac Studio (M4 Max, 64 GB): memory by quantization

Quant	Memory needed	Fits 48 GB?	Max context	Est. speed	Download
F16	81.5 GB	✗ No	—	—	70 GB
Q8_0	48.7 GB	✗ No	—	—	37.2 GB
Q6_K	40.2 GB	✓ Yes	8K	~8.6 tok/s	28.7 GB
Q5_K_M	36.3 GB	✓ Yes	16K	~9.6 tok/s	24.8 GB
Q4_K_M	32.7 GB	✓ Yes	16K	~10.9 tok/s	21.1 GB
Q3_K_M	26.5 GB	✓ Yes	16K	~14.1 tok/s	14.9 GB
Q2_K	23 GB	✓ Yes	16K	~16.8 tok/s	11.5 GB

Which Command R Family sizes fit

Variant	Needs at Q4_K_M	Fits?	Est. speed
Command R+ (104B)	65.7 GB	✗ Too large	—
Command R (35B)	32.7 GB	✓ Fits	~10.9 tok/s

What to watch out for

1 larger variant of Command R Family does not fit and would need CPU offload or different hardware.
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 M4 Max 64 GB limitations

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.
64 GB unified memory at 546 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