Head to head

GLM-5.2 vs Claude Haiku 4.5

On a blended workload of one million input tokens and three million output tokens, GLM-5.2 is 1.1× cheaper than Claude Haiku 4.5.

GLM-5.2Claude Haiku 4.5
Provider Together AI Anthropic
Input / MTok $1.4 $1
Output / MTok $4.4 $5
Cache read / MTok $0.1
Context window 262,144 200,000
Blended 1M in + 3M out $14.60 $16.00
Status ga ga

GLM-5.2 detail   Claude Haiku 4.5 detail