Mixed: zhipu/glm-5-3-flash on inferencenet

FieldBeforeAfterChange
Input0.04400.0800+81.8%
Output0.600.50-16.7%
Cache read0.04300.0250-41.9%

What this means by workload

Illustrative monthly cost of the changed field(s) alone, at a representative 10M tokens/month split per workload preset. Not a substitute for the calculator, which uses your own volume and every price field.

WorkloadBefore / monthAfter / month
Coding agent$0.72$1.01
RAG chat$1.27$1.43
Summarisation (batch)$1.00$1.22
Classification$0.55$0.88
Chat assistant$2.66$2.48
Content generation$4.89$4.16

Back to model page | All changes