The real cost to run MiMo V2.5 Pro FP4 DFlash

XiaomiMiMo · 554.3B parameters · 1M context · commercial OK

MiMo V2.5 Pro FP4 DFlash — 554.3B params (XiaomiMiMo). Auto-indexed from the Hugging Face Hub (46,372 downloads). Parameter count is exact; download size and quantizations are estimates.

What it costs to run MiMo V2.5 Pro FP4 DFlash — $0 markup

On your own machine — $0. Runs free locally if you have about 375 GB of VRAM at Q4_K_M (point Spanvero at LM Studio, Ollama, or llama.cpp).
On your own rented GPU — from $8.34/hr. 6× NVIDIA A100 80GB at the direct RunPod price (≈ $7.45 a session), $0 markup.
Via your own API key — $4.53/1M tokens (blended input + output). No GPU to manage (rough estimate for this size).

Key facts

Parameters	554.3B
Context window	1M tokens
Recommended quant	Q4_K_M
VRAM to run	~375 GB (at Q4_K_M, 16.4K context)
Download size	~333 GB
License	Commercial use OK

Open the free Spanvero advisor → for the live, interactive math for your exact workload and hardware.

Related models

MiMo 7B Base — 7.8B, XiaomiMiMo
MiMo 7B RL — 7.8B, XiaomiMiMo
MiMo V2.5 Pro — 1023.2B, XiaomiMiMo
MiMo V2 Flash — 309.8B, XiaomiMiMo
NVIDIA Nemotron 3 Ultra 550B A55B BF16 — 560.5B, nvidia

Browse: More Xiaomi (MiMo) models · All models · Compare