Compare AI models

Specifications
Parameters
236B—
Context window
128k128k
Benchmarks
HumanEval
84.76%83.7%
MATH
71.02%64.3%
Overview
CompanyDeepSeekOpenAI
Release dateJun 28 2024Nov 6 2023
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

DeepSeek V2-0628 leads GPT-4 Turbo on 2 of the 2 benchmarks they both report (HumanEval, MATH). GPT-4 Turbo shipped 235 days before DeepSeek V2-0628, so benchmark comparisons should account for the intervening progress.

Context windows are 128k (DeepSeek V2-0628) vs 128k (GPT-4 Turbo). DeepSeek V2-0628 is open weight, while GPT-4 Turbo is closed.

On HumanEval, DeepSeek V2-0628 leads at 84.76% vs GPT-4 Turbo at 83.7%. On MATH, DeepSeek V2-0628 leads at 71.02% vs GPT-4 Turbo at 64.3%.