Compare AI models

Specifications
Parameters
22B—
Context window
32k128k
Benchmarks
HumanEval
81.1%75.7%
Overview
CompanyMistralZ.ai
Release dateMay 29 2024Jun 5 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

Codestral 22B leads GLM-4-Air on 1 of the 1 benchmark they both report (HumanEval). Codestral 22B shipped 7 days before GLM-4-Air, so benchmark comparisons should account for the intervening progress.

Context windows are 32k (Codestral 22B) vs 128k (GLM-4-Air). Codestral 22B is open weight, while GLM-4-Air is closed.

On HumanEval, Codestral 22B leads at 81.1% vs GLM-4-Air at 75.7%.