During the previous 24 hours, Phala Confidential AI successfully processed 178.40 billion tokens. This operation is fully supported by a robust TEE-backed AI infrastructure. A breakdown of the most utilized models within this timeframe shows that GLM-5.3 led the way, accounting for 47.87% of the activity. The deepseek-v4-pro-0813 model followed with a 12.79% share, while GLM-5.3 Flash captured 10.08%. You can find more information at the following link: https://phala.com/models/z-ai/glm-5.3