Compare AI models

Specifications
Parameters
236B8B/70B
Context window
128k8k
Benchmarks
HumanEval
84.76%81.7%
MATH
71.02%50.4%
Overview
CompanyDeepSeekMeta
Release dateJun 28 2024Apr 18 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V2-0628 leads LLaMA 3 (8B/70B) on 2 of the 2 benchmarks they both report (HumanEval, MATH). LLaMA 3 (8B/70B) shipped 71 days before DeepSeek V2-0628, so benchmark comparisons should account for the intervening progress.

DeepSeek V2-0628 has 236B parameters, while LLaMA 3 (8B/70B) has 8B/70B. Context windows are 128k (DeepSeek V2-0628) vs 8k (LLaMA 3 (8B/70B)).

On HumanEval, DeepSeek V2-0628 leads at 84.76% vs LLaMA 3 (8B/70B) at 81.7%. On MATH, DeepSeek V2-0628 leads at 71.02% vs LLaMA 3 (8B/70B) at 50.4%.