Compare AI models

Specifications
Parameters
314B—
Context window
8k—
Benchmarks
HumanEval
63.2%72%
MATH
23.9%47.9%
MMLU
73%81.5%
GSM8K
62.9%87.6%
Overview
CompanySpaceXSIZ.ai
Release dateNov 3 2023Jan 16 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

GLM-4 leads Grok‑1 on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). Grok‑1 shipped 74 days before GLM-4, so benchmark comparisons should account for the intervening progress.

Grok‑1 is open weight, while GLM-4 is closed.

On HumanEval, GLM-4 leads at 72% vs Grok‑1 at 63.2%. On MATH, GLM-4 leads at 47.9% vs Grok‑1 at 23.9%. On MMLU, GLM-4 leads at 81.5% vs Grok‑1 at 73%. On GSM8K, GLM-4 leads at 87.6% vs Grok‑1 at 62.9%.