Compare AI models

Specifications
Parameters
7B/67B14B
Context window
4k2k
Benchmarks
HumanEval
73.8%32.3%
MATH
32.6%24.8%
MMLU
71.1%66.3%
GSM8K
84.1%61.3%
Overview
CompanyDeepSeekQwen
Release dateNov 29 2023Sep 25 2023
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek-LLM leads Qwen-14B on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). Qwen-14B shipped 65 days before DeepSeek-LLM, so benchmark comparisons should account for the intervening progress.

DeepSeek-LLM has 7B/67B parameters, while Qwen-14B has 14B. Context windows are 4k (DeepSeek-LLM) vs 2k (Qwen-14B).

On HumanEval, DeepSeek-LLM leads at 73.8% vs Qwen-14B at 32.3%. On MATH, DeepSeek-LLM leads at 32.6% vs Qwen-14B at 24.8%. On MMLU, DeepSeek-LLM leads at 71.1% vs Qwen-14B at 66.3%. On GSM8K, DeepSeek-LLM leads at 84.1% vs Qwen-14B at 61.3%.