Xiaomi/Mixture of Experts

XMiMo V2.5 Pro

Name: MiMo V2.5 Pro
Author: Xiaomi

MiMo-V2.5 is a native omnimodal model with strong agentic capabilities, supporting text, image, video, and audio understanding within a unified architecture.

chatcodingreasoningmultilingualmathvisionagentictool_use

1023.24B

Parameters (42B active)

1024K

Context length

Benchmarks

Quantizations

Architecture

MoE

Released

2026-04-22

Layers

KV Heads

Head Dim

192

Family

mimo

Quantization Options

Context length:

Quant	Bits	VRAM @ 16K	Quality
IQ2_XXS	2.38	309.8 GB 304.9 + 4.9 KV	low
IQ2_M	2.93	380.2 GB 375.3 + 4.9 KV	low
Q2_K	3.16	409.6 GB 404.7 + 4.9 KV	low
IQ3_XXS	3.25	421.1 GB 416.2 + 4.9 KV	low
IQ3_XS	3.5	453.1 GB 448.2 + 4.9 KV	low
Q3_K_S	3.64	471.0 GB 466.1 + 4.9 KV	low
IQ3_M	3.76	486.3 GB 481.4 + 4.9 KV	low
Q3_K_M	4	517.0 GB 512.1 + 4.9 KV	low
Q3_K_L	4.3	555.4 GB 550.5 + 4.9 KV	moderate
IQ4_XS	4.46	575.9 GB 570.9 + 4.9 KV	moderate
Q4_K_S	4.67	602.7 GB 597.8 + 4.9 KV	moderate
Q4_K_M	4.89	630.9 GB 625.9 + 4.9 KV	good
Q5_K_S	5.57	717.8 GB 712.9 + 4.9 KV	good
Q5_K_M	5.7	734.5 GB 729.5 + 4.9 KV	good
Q6_K	6.56	844.5 GB 839.5 + 4.9 KV	excellent
Q8_0	8.5	1092.6 GB 1087.7 + 4.9 KV	lossless
FP16	16	2051.9 GB 2047.0 + 4.9 KV	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

Too big for a single GPU — plan a multi-GPU deployment

Even the lightest quant needs ~310 GB. Size GPUs, replicas, TCO and scaling for a production setup. Open in Enterprise →

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN MIMO V2.5 PRO NOW

Rent on Vast.ai →Or RunPod →

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (9)

τ²-Bench94.2

GPQA Diamond86.6

IFBench79.9

AA Long Context73.3

AA Intelligence53.8

SciCode50.2

AA Coding45.5

Terminal-Bench43.2

HLE33.8

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

curl -fsSL https://ollama.com/install.sh | sh

$ollama run mimo:1023b-q4_K_M

Tag may need adjustment — check ollama.com/library/mimo for available tags.

▸ SETUP GUIDE

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more

Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

HuggingFace GGUF Downloads Build Hardware

Find the best GPU for MiMo V2.5 Pro

Build Hardware for MiMo V2.5 Pro

QUANT	BPW	VRAM	QUALITY
IQ2_XXS	2.38	304.9 GB	65%
IQ2_M	2.93	375.3 GB	75%
Q2_K	3.16	404.7 GB	78%
IQ3_XXS	3.25	416.2 GB	82%
IQ3_XS	3.5	448.2 GB	84%
Q3_K_S	3.64	466.1 GB	85%
IQ3_M	3.76	481.4 GB	86%
Q3_K_M	4	512.1 GB	88%
Q3_K_L	4.3	550.5 GB	90%
IQ4_XS	4.46	570.9 GB	92%
Q4_K_S	4.67	597.8 GB	93%
Q4_K_M	4.89	625.9 GB	94%
Q5_K_S	5.57	712.9 GB	96%
Q5_K_M	5.7	729.5 GB	96%
Q6_K	6.56	839.5 GB	97%
Q8_0	8.5	1087.7 GB	100%
FP16	16	2047.0 GB	100%