AI Hub
All models
LLMOpen weights

Mixtral 8x22B

Mistral AI

Updated May 21, 2026 · Benchmarks via Artificial Analysis, specs & pricing via OpenRouter · Methodology · Report an error

Specifications

Type
LLM
Access
Open weights
Released
April 10, 2024
License
Apache 2.0
Parameters
141B total / 39B active (MoE)
Context window
65,536 tokens
Input
Text
Output
Text

Benchmarks

Coding

Math

MATH
41.8

General

MMLU
77.8

Our take

A scaled-up open mixture-of-experts that pushed Apache-licensed models further up the frontier.

Compare Mixtral 8x22B with

See all Mixtral 8x22B alternatives →