Compare AI models

Specifications
Parameters
236B9B/27B
Context window
128k8k
API pricing
Cheapest input
—$0.65NextBit
Cheapest output
—$0.65NextBit
—$163NextBit · int4
Benchmarks
HumanEval
84.76%51.8%
MATH
71.02%42.3%
Overview
CompanyDeepSeekGoogle
Release dateJun 28 2024Jun 27 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V2-0628 leads Gemma 2 on 2 of the 2 benchmarks they both report (HumanEval, MATH). Gemma 2 shipped 1 day before DeepSeek V2-0628, so benchmark comparisons should account for the intervening progress.

DeepSeek V2-0628 has 236B parameters, while Gemma 2 has 9B/27B. Context windows are 128k (DeepSeek V2-0628) vs 8k (Gemma 2).

On HumanEval, DeepSeek V2-0628 leads at 84.76% vs Gemma 2 at 51.8%. On MATH, DeepSeek V2-0628 leads at 71.02% vs Gemma 2 at 42.3%.