Compare AI models

Specifications
Parameters
236B3B
Context window
128k128k
Benchmarks
HumanEval
84.76%77.4%
MATH
71.02%51.7%
Overview
CompanyDeepSeekMistral
Release dateJun 28 2024Oct 16 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

DeepSeek V2-0628 leads Ministral 3B on 2 of the 2 benchmarks they both report (HumanEval, MATH). DeepSeek V2-0628 shipped 110 days before Ministral 3B, so benchmark comparisons should account for the intervening progress.

DeepSeek V2-0628 has 236B parameters, while Ministral 3B has 3B. Context windows are 128k (DeepSeek V2-0628) vs 128k (Ministral 3B). DeepSeek V2-0628 is open weight, while Ministral 3B is closed.

On HumanEval, DeepSeek V2-0628 leads at 84.76% vs Ministral 3B at 77.4%. On MATH, DeepSeek V2-0628 leads at 71.02% vs Ministral 3B at 51.7%.