Compare AI models

Specifications
Parameters
—14.3B
Context window
32k32k
Benchmarks
HumanEval
45.1%34.2%
MMLU
81.2%62.5%
GSM8K
81%61.5%
Overview
CompanyMistralQwen
Release dateFeb 26 2024Mar 28 2024
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Mistral Large leads Qwen1.5-MoE on 3 of the 3 benchmarks they both report (HumanEval, MMLU, GSM8K). Mistral Large shipped 31 days before Qwen1.5-MoE, so benchmark comparisons should account for the intervening progress.

Context windows are 32k (Mistral Large) vs 32k (Qwen1.5-MoE). Mistral Large is closed, while Qwen1.5-MoE is open weight.

On HumanEval, Mistral Large leads at 45.1% vs Qwen1.5-MoE at 34.2%. On MMLU, Mistral Large leads at 81.2% vs Qwen1.5-MoE at 62.5%. On GSM8K, Mistral Large leads at 81% vs Qwen1.5-MoE at 61.5%.