Compare AI models

Specifications
Parameters
141B14.3B
Context window
64k32k
API pricing
Cheapest input
$2.00Mistral—
Cheapest output
$6.00Mistral—
$700Mistral · zdr—
Benchmarks
HumanEval
45.1%34.2%
MMLU
77.75%62.5%
GSM8K
78.6%61.5%
Overview
CompanyMistralQwen
Release dateApr 10 2024Mar 28 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Mixtral 8×22B leads Qwen1.5-MoE on 3 of the 3 benchmarks they both report (HumanEval, MMLU, GSM8K). Qwen1.5-MoE shipped 13 days before Mixtral 8×22B, so benchmark comparisons should account for the intervening progress.

Mixtral 8×22B has 141B parameters, while Qwen1.5-MoE has 14.3B. Context windows are 64k (Mixtral 8×22B) vs 32k (Qwen1.5-MoE).

On HumanEval, Mixtral 8×22B leads at 45.1% vs Qwen1.5-MoE at 34.2%. On MMLU, Mixtral 8×22B leads at 77.75% vs Qwen1.5-MoE at 62.5%. On GSM8K, Mixtral 8×22B leads at 78.6% vs Qwen1.5-MoE at 61.5%.