Compare AI models

Specifications
Parameters
16B7B
Context window
32k256k
Benchmarks
HumanEval
57.3%75%
Overview
CompanyDeepSeekMistral
Release dateMay 16 2024Jul 16 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Codestral Mamba 7B leads DeepSeek V2 Lite on 1 of the 1 benchmark they both report (HumanEval). DeepSeek V2 Lite shipped 61 days before Codestral Mamba 7B, so benchmark comparisons should account for the intervening progress.

DeepSeek V2 Lite has 16B parameters, while Codestral Mamba 7B has 7B. Context windows are 32k (DeepSeek V2 Lite) vs 256k (Codestral Mamba 7B).

On HumanEval, Codestral Mamba 7B leads at 75% vs DeepSeek V2 Lite at 57.3%.