Compare AI models

Specifications
Parameters
7B/67B14.3B
Context window
4k32k
Benchmarks
HumanEval
73.8%34.2%
MMLU
71.1%62.5%
GSM8K
84.1%61.5%
Overview
CompanyDeepSeekQwen
Release dateNov 29 2023Mar 28 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek-LLM leads Qwen1.5-MoE on 3 of the 3 benchmarks they both report (HumanEval, MMLU, GSM8K). DeepSeek-LLM shipped 120 days before Qwen1.5-MoE, so benchmark comparisons should account for the intervening progress.

DeepSeek-LLM has 7B/67B parameters, while Qwen1.5-MoE has 14.3B. Context windows are 4k (DeepSeek-LLM) vs 32k (Qwen1.5-MoE).

On HumanEval, DeepSeek-LLM leads at 73.8% vs Qwen1.5-MoE at 34.2%. On MMLU, DeepSeek-LLM leads at 71.1% vs Qwen1.5-MoE at 62.5%. On GSM8K, DeepSeek-LLM leads at 84.1% vs Qwen1.5-MoE at 61.5%.