Compare AI models

Specifications
Parameters
—6B
Context window
8k32k
API pricing
Cheapest input
$30.00Azure—
Cheapest output
$60.00Azure—
$9,000Azure—
Benchmarks
HumanEval
67%9.8%
HellaSwag
95.3%57%
MMLU
86.4%45.2%
GSM8K
92%25.9%
Overview
CompanyOpenAIZ.ai
Release dateMar 14 2023Jun 25 2023
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

GPT-4 leads ChatGLM2-6B on 4 of the 4 benchmarks they both report (HumanEval, HellaSwag, MMLU, GSM8K). GPT-4 shipped 103 days before ChatGLM2-6B, so benchmark comparisons should account for the intervening progress.

Context windows are 8k (GPT-4) vs 32k (ChatGLM2-6B). GPT-4 is closed, while ChatGLM2-6B is open weight.

On HumanEval, GPT-4 leads at 67% vs ChatGLM2-6B at 9.8%. On HellaSwag, GPT-4 leads at 95.3% vs ChatGLM2-6B at 57%. On MMLU, GPT-4 leads at 86.4% vs ChatGLM2-6B at 45.2%. On GSM8K, GPT-4 leads at 92% vs ChatGLM2-6B at 25.9%.