Compare AI models

Specifications
Parameters
7B/13B/33B/65B—
Context window
2k—
Benchmarks
HumanEval
23.7%72%
MATH
10.6%47.9%
MMLU
63.4%81.5%
GSM8K
50.9%87.6%
Overview
CompanyMetaZ.ai
Release dateFeb 24 2023Jan 16 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

GLM-4 leads LLaMA 1 on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). LLaMA 1 shipped 326 days before GLM-4, so benchmark comparisons should account for the intervening progress.

LLaMA 1 is open weight, while GLM-4 is closed.

On HumanEval, GLM-4 leads at 72% vs LLaMA 1 at 23.7%. On MATH, GLM-4 leads at 47.9% vs LLaMA 1 at 10.6%. On MMLU, GLM-4 leads at 81.5% vs LLaMA 1 at 63.4%. On GSM8K, GLM-4 leads at 87.6% vs LLaMA 1 at 50.9%.