Compare AI models

Specifications
Parameters
123B14.3B
Context window
128k32k
API pricing
Cheapest input
$2.00Mistral—
Cheapest output
$6.00Mistral—
$700Mistral—
Benchmarks
HumanEval
92.1%34.2%
MMLU
84%62.5%
Overview
CompanyMistralQwen
Release dateJul 24 2024Mar 28 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Mistral Large 2 leads Qwen1.5-MoE on 2 of the 2 benchmarks they both report (HumanEval, MMLU). Qwen1.5-MoE shipped 118 days before Mistral Large 2, so benchmark comparisons should account for the intervening progress.

Mistral Large 2 has 123B parameters, while Qwen1.5-MoE has 14.3B. Context windows are 128k (Mistral Large 2) vs 32k (Qwen1.5-MoE).

On HumanEval, Mistral Large 2 leads at 92.1% vs Qwen1.5-MoE at 34.2%. On MMLU, Mistral Large 2 leads at 84% vs Qwen1.5-MoE at 62.5%.