What are the hardware requirements for Mixtral 8x22B Instruct V0.1?

Mixtral 8x22B Instruct V0.1 requires approximately 221.4 GB of VRAM and supports a context window of 64K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Mixtral 8x22B Instruct V0.1 and how large is it?

Mixtral 8x22B Instruct V0.1 is developed by mistralai, a model with 140.6b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Mixtral 8x22B Instruct V0.1 perform on standard benchmarks?

Mixtral 8x22B Instruct V0.1 has the following published scores: MMLU 77.77, LMSYS ELO 1229. Compare against reference models on this page or on the LLM Explorer leaderboards.

Mixtral 8x22B Instruct V0.1 by mistralai — VRAM 221.4GB, 64K context

Name: Mixtral 8x22B Instruct V0.1
Author: mistralai

Mixtral 8x22B Instruct V0.1 is an open-source language model by mistralai. Features: 140.6b LLM, VRAM: 221.4GB, Context: 64K, License: apache-2.0, MoE, Instruction-Based, HF Score: 79.2, LLM Explorer Score: 0.43, ELO: 1229, Arc: 72.7, HellaSwag: 89.1, MMLU: 77.8, TruthfulQA: 68.1, WinoGrande: 85.2, GSM8K: 82.

Base model:finetune:mistralai/... Base model:mistralai/mixtral-8... De Deploy:azure En Es Fr Instruct It Mistral-common Mixtral Moe Region:us Safetensors Sharded Tensorflow Vllm

Model Card on HF 🤗: https://huggingface.co/mistralai/Mixtral-8x22B-Instruct-v0.1

Mixtral 8x22B Instruct V0.1 Benchmarks

LMSys ELO: 1229 vs 1272 (so35)^-3.4%

IFEval: 71.84 vs 88 (so35)^-18.4%

ARC: 72.7 vs 96.7 (so35)^-24.8%

HellaSwag: 89.08 vs 95.3 (gpt4)^-6.5%

MMLU: 77.77 vs 88.3 (so35)^-11.9%

TruthfulQA: 68.14 vs 59 (gpt4)^15.5%

WinoGrande: 85.16 vs 87.5 (gpt4)^-2.7%

GSM8K: 82.03 vs 96.4 (so35)^-14.9%

MATH Lvl 5: 18.73

LLME Score: 0.43084

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Mixtral 8x22B Instruct V0.1 (mistralai/Mixtral-8x22B-Instruct-v0.1)

🌟 Advertise your project 🚀

Mixtral 8x22B Instruct V0.1 Parameters and Internals

Model Type

text generation, instruction following

Additional Notes

The HuggingFace tokenizer included in this release should match the Mistral tokenizer, ensuring consistent results.

Supported Languages

en (.), es (.), it (.), de (.), fr (.)

LLM Name	Mixtral 8x22B Instruct V0.1
Repository 🤗	https://huggingface.co/mistralai/Mixtral-8x22B-Instruct-v0.1
Base Model(s)	mistralai/Mixtral-8x22B-v0.1 mistralai/Mixtral-8x22B-v0.1
Model Size	140.6b
Required VRAM	221.4 GB
Updated	2026-04-14
Maintainer	mistralai
Model Type	mixtral
Instruction-Based	Yes
Model Files	4.8 GB: 1-of-59 4.8 GB: 2-of-59 4.8 GB: 3-of-59 4.8 GB: 4-of-59 4.8 GB: 5-of-59 4.8 GB: 6-of-59 4.8 GB: 7-of-59 4.8 GB: 8-of-59 4.8 GB: 9-of-59 4.8 GB: 10-of-59 4.8 GB: 11-of-59 4.8 GB: 12-of-59 4.8 GB: 13-of-59 4.8 GB: 14-of-59 4.8 GB: 15-of-59 4.8 GB: 16-of-59 4.8 GB: 17-of-59 4.8 GB: 18-of-59 4.8 GB: 19-of-59 4.8 GB: 20-of-59 4.8 GB: 21-of-59 4.8 GB: 22-of-59 4.9 GB: 23-of-59 5.0 GB: 24-of-59 5.0 GB: 25-of-59 4.9 GB: 26-of-59 4.8 GB: 27-of-59 4.8 GB: 28-of-59 4.8 GB: 29-of-59 4.8 GB: 30-of-59 4.8 GB: 31-of-59 4.8 GB: 32-of-59 4.8 GB: 33-of-59 4.8 GB: 34-of-59 4.8 GB: 35-of-59 4.8 GB: 36-of-59 4.8 GB: 37-of-59 4.8 GB: 38-of-59 4.8 GB: 39-of-59 4.8 GB: 40-of-59 4.8 GB: 41-of-59 4.8 GB: 42-of-59 4.8 GB: 43-of-59 4.8 GB: 44-of-59 4.8 GB: 45-of-59 4.8 GB: 46-of-59
Supported Languages	en es it de fr
Model Architecture	MixtralForCausalLM
License	apache-2.0
Context Length	65536
Model Max Length	65536
Transformers Version	4.38.0
Vocabulary Size	32768
Torch Data Type	bfloat16

Quantized Models of the Mixtral 8x22B Instruct V0.1

Model	Likes	Downloads	VRAM
...ixtral 8x22B Instruct V0.1 AWQ	13	388	73 GB

Best Alternatives to Mixtral 8x22B Instruct V0.1

Best Alternatives	Context / RAM	Downloads	Likes
...ixtral 8x22B Instruct V0.1 FP8	64K / 140.9 GB	548	0
...igHuggyD Grey WizardLM 2 8x22B	64K / 216.6 GB	6	4
...x22B Instruct V0.1 FP8 Dynamic	64K / 140.9 GB	22	0
WizardLM 2 8x22B Beige	64K / 221.4 GB	3	4
...ral 8x22B Instruct V0.1 FP8 V2	64K / 140.9 GB	5	0
...ral 8x22B Instruct V0.1 FP8 V1	64K / 140.9 GB	8	0
Mixtral 8x22B GO Instruct V1	64K / 280.8 GB	5	2
Mixtral 8x22B Instruct V0.1	64K / 221.4 GB	17	1
Mixtral 8x22b Instruct Oh	64K / 221.6 GB	10	29
... 8x22B V0.1 Instruct Sft En De	64K / 212 GB	12	1

Rank the Mixtral 8x22B Instruct V0.1 Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

What open-source LLMs or SLMs are you in search of? 53254 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Check out Ag3ntum — our secure, self-hosted AI agent for server management.

Release v20260328a

Support LLM Explorer

Mixtral 8x22B Instruct V0.1 by mistralai