Compare AI models

Specifications
Parameters
16B—
Context window
8k—
Benchmarks
HumanEval
48.1%72%
MATH
45.3%47.9%
MMLU
70%81.5%
GSM8K
77.4%87.6%
Overview
CompanyMoonshot AIZ.ai
Release dateFeb 22 2025Jan 16 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

GLM-4 leads Moonlight on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). GLM-4 shipped 403 days before Moonlight, so benchmark comparisons should account for the intervening progress.

Moonlight is open weight, while GLM-4 is closed.

On HumanEval, GLM-4 leads at 72% vs Moonlight at 48.1%. On MATH, GLM-4 leads at 47.9% vs Moonlight at 45.3%. On MMLU, GLM-4 leads at 81.5% vs Moonlight at 70%. On GSM8K, GLM-4 leads at 87.6% vs Moonlight at 77.4%.