Compare AI models

Specifications
Parameters
236B14B
Context window
128k2k
Benchmarks
HumanEval
81.1%32.3%
MATH
53.9%24.8%
MMLU
77.8%66.3%
GSM8K
92.2%61.3%
Overview
CompanyDeepSeekQwen
Release dateMay 6 2024Sep 25 2023
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V2 leads Qwen-14B on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). Qwen-14B shipped 224 days before DeepSeek V2, so benchmark comparisons should account for the intervening progress.

DeepSeek V2 has 236B parameters, while Qwen-14B has 14B. Context windows are 128k (DeepSeek V2) vs 2k (Qwen-14B).

On HumanEval, DeepSeek V2 leads at 81.1% vs Qwen-14B at 32.3%. On MATH, DeepSeek V2 leads at 53.9% vs Qwen-14B at 24.8%. On MMLU, DeepSeek V2 leads at 77.8% vs Qwen-14B at 66.3%. On GSM8K, DeepSeek V2 leads at 92.2% vs Qwen-14B at 61.3%.