Compare AI models

Specifications
Parameters
16B6B
Context window
32k32k
Benchmarks
HumanEval
57.3%9.8%
MATH
27.9%6.9%
MMLU
55.7%45.2%
GSM8K
72%25.9%
Overview
CompanyDeepSeekZ.ai
Release dateMay 16 2024Jun 25 2023
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V2 Lite leads ChatGLM2-6B on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). ChatGLM2-6B shipped 326 days before DeepSeek V2 Lite, so benchmark comparisons should account for the intervening progress.

DeepSeek V2 Lite has 16B parameters, while ChatGLM2-6B has 6B. Context windows are 32k (DeepSeek V2 Lite) vs 32k (ChatGLM2-6B).

On HumanEval, DeepSeek V2 Lite leads at 57.3% vs ChatGLM2-6B at 9.8%. On MATH, DeepSeek V2 Lite leads at 27.9% vs ChatGLM2-6B at 6.9%. On MMLU, DeepSeek V2 Lite leads at 55.7% vs ChatGLM2-6B at 45.2%. On GSM8K, DeepSeek V2 Lite leads at 72% vs ChatGLM2-6B at 25.9%.