Mixtral 8x22B

Mixtral 8x22B is a powerful sparse Mixture of Experts (MoE) model with 141B total parameters and 39B active per token. Features a 64K context window, exceptional math performance, and cost-efficient inference. Supports English, French, German, Spanish, and Italian. Apache 2.0 licensed.

Added Dec 11, 2025

Model weights

Pricing

Auto routing · per 1M tokens

Input: $2.00
Output: $6.00
Cache read: $0.20

Compare provider prices

Specifications

Context window: 65.5K
Max output: 65.5K
Parameters: 141B / 39B; Total / active
Avg output (7d): 300 tokens; Longer than 25% of models

Benchmarks

Artificial Analysis

LMArena

Sourced from Artificial Analysis.

Intelligence Index

5.7

Better than 13% of models compared

Reasoning

HLE

Humanity's Last Exam

4.0%

Better than 12% of models compared

Coding

LiveCodeBench

Contamination-free coding benchmark

14.8%

Better than 15% of models compared

Math

AIME

American Invitational Mathematics Examination

0.0%

Better than 3% of models compared

Math-500

Diverse mathematical problem solving benchmark

54.5%

Better than 17% of models compared

Knowledge

MMLU-Pro

Professional and academic subject knowledge

53.7%

Better than 17% of models compared

Legacy benchmarks

GPQA Diamond (legacy)

Graduate-level scientific reasoning

33.2%

Better than 10% of models compared

Last updated Oct 11, 2026