GLM-5.3 Announces Safety Evaluation Progress, Open Weights Release Delayed
Z.ai announces progress on the safety evaluation of the new generation GLM-5.3; due to the model's improved capabilities in cybersecurity tasks, the release of open weights will be delayed to complete further testing and protection.
GLM-5.3 is the latest cutting-edge model announced by Z.ai. At this stage, the most important information is not unverified parameters, but the vendor's adjustment to the model's safety release schedule.
According to public reports, Z.ai has delayed the release of open weights by two weeks to conduct additional safety and protection testing. The page has therefore marked it as "Under Safety Evaluation" rather than "Fully Open".
KEY HIGHLIGHTS
Key highlights
Designed for complex reasoning and agent tasks
Enhanced cybersecurity capability evaluation
Open weights will be released after safety testing