Can I Run MiMo-V2.5 on 256 GB system RAM?

Autor: Jakub Rusinowski · Ostatnia aktualizacja: 6 września 2026

Yes, but it is tight

Yes, but it is tight — MiMo-V2.5 310B at Q4_K_M needs about 191.9 GB of the 204.8 GB usable on 256 GB system RAM, leaving only ~12.9 GB before the runtime starts swapping. Expect ~6 tok/s (estimated), with room for about 32,768 tokens of context.

Confidence: low · Recommended quantization: Q4_K_M · Estimated speed: ~6 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.

256 GB system RAM — what it gives a model

Usable memory for models	204.8 GB
Memory bandwidth	90 GB/s

MiMo-V2.5 on 256 GB system RAM: memory by quantization

Quant	Memory needed	Fits 204.8 GB?	Max context	Est. speed	Download
F16	624.7 GB	✗ No	—	—	620 GB
Q8_0	334.1 GB	✗ No	—	—	329.4 GB
Q6_K	258.9 GB	✗ No	—	—	254.2 GB
Q5_K_M	224.4 GB	✗ No	—	—	219.7 GB
Q4_K_M	191.9 GB	✓ Yes	32K	~6 tok/s	187.2 GB
Q3_K_M	136.8 GB	✓ Yes	128K	~7.9 tok/s	132.1 GB
Q2_K	106.6 GB	✓ Yes	128K	~9.6 tok/s	101.9 GB

Which MiMo-V2.5 sizes fit

Variant	Needs at Q4_K_M	Fits?	Est. speed
MiMo-V2.5-Pro 1T	610.3 GB	✗ Too large	—
MiMo-V2.5 310B	191.9 GB	✓ Fits	~6 tok/s

What to watch out for

1 larger variant of MiMo-V2.5 does not fit and would need CPU offload or different hardware.
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.
204.8 GB of the 256 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