Can I Run MiniMax M3 on 192 GB system RAM?

作者： Jakub Rusinowski · 最后更新： 2026年6月26日

Yes, but it is tight

Yes, but it is tight — MiniMax M3 230B-A10B at Q4_K_M needs about 143.2 GB of the 153.6 GB usable on 192 GB system RAM, leaving only ~10.4 GB before the runtime starts swapping. Expect ~8.5 tok/s (estimated), with room for about 16,384 tokens of context.

Confidence: low · Recommended quantization: Q4_K_M · Estimated speed: ~8.5 tok/s

Buy This HardwareApple Mac Studio M2 Ultra — 192 GB VRAM · 60 W board power Deploy in the Cloud NowRunPod

or compare on Vast.ai

As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.

192 GB system RAM — what it gives a model

Usable memory for models	153.6 GB
Memory bandwidth	90 GB/s

MiniMax M3 on 192 GB system RAM: memory by quantization

Quant	Memory needed	Fits 153.6 GB?	Max context	Est. speed	Download
F16	464.3 GB	✗ No	—	—	460 GB
Q8_0	248.7 GB	✗ No	—	—	244.4 GB
Q6_K	192.9 GB	✗ No	—	—	188.6 GB
Q5_K_M	167.3 GB	✗ No	—	—	163 GB
Q4_K_M	143.2 GB	✓ Yes	16K	~8.5 tok/s	138.9 GB
Q3_K_M	102.4 GB	✓ Yes	64K	~10.9 tok/s	98 GB
Q2_K	79.9 GB	✓ Yes	128K	~12.9 tok/s	75.6 GB

Which MiniMax M3 sizes fit

Variant	Needs at Q4_K_M	Fits?	Est. speed
MiniMax M3 230B-A10B	143.2 GB	✓ Fits	~8.5 tok/s
MiniMax M3-VL	143.2 GB	✓ Fits	~8.5 tok/s

What to watch out for

This model does not publish its full attention configuration, so the KV-cache share of these figures is inferred from its parameter count rather than computed exactly.
These figures assume CPU-only inference. Any discrete GPU, even an 8 GB one, will be several times faster for models that fit in its VRAM.

Recommended setup

llama.cpp (CPU build) or Ollama — both run without a GPU

How these numbers are calculated

Memory need = quantized weights + KV cache at 8,192 tokens (f16) + 0.8 GB runtime overhead.
153.6 GB of the 192 GB is treated as usable for model weights (80% — the rest is the OS and running applications).
DDR5-5600 dual channel at 89.6 GB/s peak. CPU decode is assumed to sustain 35% of that peak, because CPU inference is not purely bandwidth-bound — it also spends real time in compute and thread synchronisation. This figure is an assumption, not a fitted constant: no CPU measurement is in the calibration set.
CPU-only inference: no GPU is assumed. A GPU of any size will beat these figures substantially.
Throughput is a memory-bandwidth roofline estimate, not a measurement. It is labelled "estimated" everywhere it appears.
KV cache is inferred from the parameter count because this model does not publish its full attention configuration.

FAQ