Compare AI models

Specifications
Parameters
7B/13B/70B—
Context window
4k128k
Benchmarks
HumanEval
29.9%75.7%
MATH
13.5%57.9%
MMLU
68.9%81.9%
GSM8K
56.8%90.9%
Overview
CompanyMetaZ.ai
Release dateJul 18 2023Jun 5 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

GLM-4-Air leads LLaMA 2 on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). LLaMA 2 shipped 323 days before GLM-4-Air, so benchmark comparisons should account for the intervening progress.

Context windows are 4k (LLaMA 2) vs 128k (GLM-4-Air). LLaMA 2 is open weight, while GLM-4-Air is closed.

On HumanEval, GLM-4-Air leads at 75.7% vs LLaMA 2 at 29.9%. On MATH, GLM-4-Air leads at 57.9% vs LLaMA 2 at 13.5%. On MMLU, GLM-4-Air leads at 81.9% vs LLaMA 2 at 68.9%. On GSM8K, GLM-4-Air leads at 90.9% vs LLaMA 2 at 56.8%.