Compare AI models

Specifications
Parameters
236B—
Context window
128k200k
Benchmarks
HumanEval
87.2%99.3%
Overview
CompanyDeepSeekOpenAI
Release dateJul 24 2024Apr 16 2025
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

o4-mini-high leads DeepSeek Coder V2-0724 on 1 of the 1 benchmark they both report (HumanEval). DeepSeek Coder V2-0724 shipped 266 days before o4-mini-high, so benchmark comparisons should account for the intervening progress.

Context windows are 128k (DeepSeek Coder V2-0724) vs 200k (o4-mini-high). DeepSeek Coder V2-0724 is open weight, while o4-mini-high is closed.

On HumanEval, o4-mini-high leads at 99.3% vs DeepSeek Coder V2-0724 at 87.2%.