Xiaomi/Mixture of Experts

XMiMo V2.6 Pro

Name: MiMo V2.6 Pro
Author: Xiaomi

MiMo V2.6 Pro is Xiaomi's 1.02T omnimodal MoE flagship with 42B active, 1M context and sliding-window attention on 60 of 70 layers. Xiaomi's API now routes MiMo-V2.5-Pro traffic to it.

chatcodingreasoningmultilingualvisionaudiomathagentictool_use

1024.22B

Parameters (42B active)

1024K

Context length

Benchmarks

Quantizations

28K

HF downloads

Architecture

MoE

Released

2026-09-21

Layers

KV Heads

Head Dim

192

Family

mimo

Quantization Options

Context length:

Quant	Bits	VRAM @ 16K	Quality
IQ2_XXS	2.38	305.9 GB 305.2 + 0.7 KV	low
IQ2_M	2.93	376.4 GB 375.6 + 0.7 KV	low
Q2_K	3.16	405.8 GB 405.1 + 0.7 KV	low
IQ3_XXS	3.25	417.3 GB 416.6 + 0.7 KV	low
IQ3_XS	3.5	449.3 GB 448.6 + 0.7 KV	low
Q3_K_S	3.64	467.3 GB 466.5 + 0.7 KV	low
IQ3_M	3.76	482.6 GB 481.9 + 0.7 KV	low
Q3_K_M	4	513.3 GB 512.6 + 0.7 KV	low
Q3_K_L	4.3	551.8 GB 551.0 + 0.7 KV	moderate
IQ4_XS	4.46	572.2 GB 571.5 + 0.7 KV	moderate
Q4_K_S	4.67	599.1 GB 598.4 + 0.7 KV	moderate
Q4_K_M	4.89	627.3 GB 626.5 + 0.7 KV	good
Q5_K_S	5.57	714.3 GB 713.6 + 0.7 KV	good
Q5_K_M	5.7	731.0 GB 730.2 + 0.7 KV	good
Q6_K	6.56	841.1 GB 840.3 + 0.7 KV	excellent
Q8_0	8.5	1089.5 GB 1088.7 + 0.7 KV	lossless
FP16	16	2049.7 GB 2048.9 + 0.7 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 ~306 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 V2.6 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 (4)

AA Long Context86.3

SciCode60.9

HLE49.4

Terminal-Bench 434.8

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

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

$ollama run mimo:1024b-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 V2.6 Pro

Build Hardware for MiMo V2.6 Pro

QUANT	BPW	VRAM	QUALITY
IQ2_XXS	2.38	305.2 GB	65%
IQ2_M	2.93	375.6 GB	75%
Q2_K	3.16	405.1 GB	78%
IQ3_XXS	3.25	416.6 GB	82%
IQ3_XS	3.5	448.6 GB	84%
Q3_K_S	3.64	466.5 GB	85%
IQ3_M	3.76	481.9 GB	86%
Q3_K_M	4	512.6 GB	88%
Q3_K_L	4.3	551.0 GB	90%
IQ4_XS	4.46	571.5 GB	92%
Q4_K_S	4.67	598.4 GB	93%
Q4_K_M	4.89	626.5 GB	94%
Q5_K_S	5.57	713.6 GB	96%
Q5_K_M	5.7	730.2 GB	96%
Q6_K	6.56	840.3 GB	97%
Q8_0	8.5	1088.7 GB	100%
FP16	16	2048.9 GB	100%