Zhipu AI's Z.ai Platform Reveals 19x Token Billing Anomaly in Caching System

According to Beating, Zhipu AI's overseas developer platform Z.ai has revealed a significant caching billing system anomaly. Developer Da7em reported on X that repeated context inputs were not being correctly charged as cached tokens but instead billed as new inputs. Da7em's billing screenshot showed the development tool Zcode recorded approximately 270,000 tokens in actual usage, while the platform's final invoice displayed nearly 5 million tokens consumed—a roughly 19-fold discrepancy. After testing multiple intelligent agent frameworks, Da7em confirmed the anomaly persisted consistently, indicating the fault originates from the server-side caching and accounting system rather than client-side tools.
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments