Compare AI models

Specifications
Context window
200k128k
Benchmarks
HumanEval
75.9%75.7%
MMLU
76.7%81.9%
GSM8K
88.9%90.9%
Overview
CompanyAnthropicZ.ai
Release dateMar 4 2024Jun 5 2024
AccessClosedClosed
Model detailsView modelView model

Frequently asked questions

GLM-4-Air leads Claude 3 Haiku on 2 of the 3 benchmarks they both report (HumanEval, MMLU, GSM8K). Claude 3 Haiku shipped 93 days before GLM-4-Air, so benchmark comparisons should account for the intervening progress.

Context windows are 200k (Claude 3 Haiku) vs 128k (GLM-4-Air).

On HumanEval, Claude 3 Haiku leads at 75.9% vs GLM-4-Air at 75.7%. On MMLU, GLM-4-Air leads at 81.9% vs Claude 3 Haiku at 76.7%. On GSM8K, GLM-4-Air leads at 90.9% vs Claude 3 Haiku at 88.9%.