Compare AI models

Specifications
Parameters
7B/67B—
Context window
4k128k
Benchmarks
HumanEval
73.8%75.7%
MATH
32.6%57.9%
MMLU
71.1%81.9%
GSM8K
84.1%90.9%
Overview
CompanyDeepSeekZ.ai
Release dateNov 29 2023Jun 5 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

GLM-4-Air leads DeepSeek-LLM on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). DeepSeek-LLM shipped 189 days before GLM-4-Air, so benchmark comparisons should account for the intervening progress.

Context windows are 4k (DeepSeek-LLM) vs 128k (GLM-4-Air). DeepSeek-LLM is open weight, while GLM-4-Air is closed.

On HumanEval, GLM-4-Air leads at 75.7% vs DeepSeek-LLM at 73.8%. On MATH, GLM-4-Air leads at 57.9% vs DeepSeek-LLM at 32.6%. On MMLU, GLM-4-Air leads at 81.9% vs DeepSeek-LLM at 71.1%. On GSM8K, GLM-4-Air leads at 90.9% vs DeepSeek-LLM at 84.1%.