Compare AI models

Specifications
Parameters
0.5B/1.5B/3B/7B/14B/32B/72B—
Context window
128k—
API pricing
Cheapest input
$0.288DeepInfra—
Cheapest output
$0.32DeepInfra—
$92.00DeepInfra · fp8—
Benchmarks
HumanEval
86.6%72%
MATH
83.1%47.9%
GSM8K
95.8%87.6%
Overview
CompanyQwenZ.ai
Release dateSep 19 2024Jan 16 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

Qwen2.5 leads GLM-4 on 3 of the 3 benchmarks they both report (HumanEval, MATH, GSM8K). GLM-4 shipped 247 days before Qwen2.5, so benchmark comparisons should account for the intervening progress.

Qwen2.5 is open weight, while GLM-4 is closed.

On HumanEval, Qwen2.5 leads at 86.6% vs GLM-4 at 72%. On MATH, Qwen2.5 leads at 83.1% vs GLM-4 at 47.9%. On GSM8K, Qwen2.5 leads at 95.8% vs GLM-4 at 87.6%.