Zhipu AI Offers Free Nighttime Usage of GLM-5.3-Flash Model
Zhipu AI has launched a promotion from September 3 to September 20, offering free nighttime usage of its GLM-5.3-Flash model for GLM Coding Plan subscribers. During off-peak hours (23:00 to 09:00 UTC+8), users incur zero quota consumption when accessing the model via the official ZCode programming tool and get double quota allowances when using other supported AI agents. This off-peak promotion incentivizes developers to test and adopt Zhipu AI's native coding environment and lightweight multimodal model for intensive development tasks. By offering zero-cost compute during night hours, Zhipu AI aims to strengthen developer engagement while balancing server compute load. The promotion applies exclusively to GLM-5.3-Flash—a native multimodal model supporting a 1-million-token context window—while usage of standard GLM-5.3 continues to consume normal quota. GLM-5.3-Flash previously debuted on OpenRouter under the pseudonym "Ox Alpha" and registered a score of 57 on the Artificial Analysis Intelligence Index.
## BACKGROUND
ZCode is Zhipu AI's desktop tool built to integrate AI coding models directly into developer workflows. OpenRouter is a aggregate platform that lets developers access and benchmark hundreds of AI models through a unified API, while the Artificial Analysis Intelligence Index evaluates model performance across reasoning, coding, and multi-step tasks.