Compare AI models

Specifications
Parameters
47B1.5B/7B/72B
Context window
32k4k
Benchmarks
GSM8K
58.4%95.9%
Overview
CompanyMistralQwen
Release dateDec 8 2023Sep 19 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Qwen2.5-Math leads Mixtral 8×7B on 1 of the 1 benchmark they both report (GSM8K). Mixtral 8×7B shipped 286 days before Qwen2.5-Math, so benchmark comparisons should account for the intervening progress.

Mixtral 8×7B has 47B parameters, while Qwen2.5-Math has 1.5B/7B/72B. Context windows are 32k (Mixtral 8×7B) vs 4k (Qwen2.5-Math).

On GSM8K, Qwen2.5-Math leads at 95.9% vs Mixtral 8×7B at 58.4%.