WizardLM/Mixture of Experts

WizardLM 2 8x22B

Name: WizardLM 2 8x22B
Author: WizardLM

WizardLM 2 8x22B — Microsoft's MoE model for complex instructions.

chatreasoning

141B

Parameters (39B active)

64K

Context length

Benchmarks

Quantizations

Architecture

MoE

Released

2024-04-15

Layers

KV Heads

Head Dim

128

Family

mistral

Quantization Options

Quant	Bits	VRAM	Quality
IQ2_XXS	2.38	42.4 GB	low
IQ2_M	2.93	52.1 GB	low
Q2_K	3.16	56.2 GB	low
IQ3_XXS	3.25	57.8 GB	low
IQ3_XS	3.5	62.2 GB	low
Q3_K_S	3.64	64.6 GB	low
IQ3_M	3.76	66.8 GB	low
Q3_K_M	4	71.0 GB	low
Q3_K_L	4.3	76.3 GB	moderate
IQ4_XS	4.46	79.1 GB	moderate
Q4_K_S	4.67	82.8 GB	moderate
Q4_K_M	4.89	86.7 GB	good
Q5_K_S	5.57	98.7 GB	good
Q5_K_M	5.7	101.0 GB	good
Q6_K	6.56	116.1 GB	excellent
Q8_0	8.5	150.3 GB	lossless
FP16	16	282.5 GB	lossless

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

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN WIZARDLM 2 8X22B NOW

Rent on RunPod →Or Vast.ai →

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

Community Ratings

Loading ratings...

Benchmarks (7)

IFEval84.0

BBH52.7

MMLU-PRO50.7

MATH49.5

GPQA24.9

GPQA Diamond17.6

MUSR17.2

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

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

$ollama run wizardlm2:8x22b-q4_K_M

Downloads and runs automatically. Add --verbose for speed stats.

▸ 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 Ollama Library GGUF Downloads Build Hardware