Powered by a robust TEE-backed AI infrastructure, Phala Confidential AI successfully processed a total of 186.42 billion tokens over the past 24 hours. A closer look at the top models reveals that GLM-5.3 Flash led the activity with a 69.41% share. The deepseek-v4.1-flash-1017 model followed with 7.61%, while xing4.0-29b-a4b accounted for 5.50% of the volume. You can find more information about the leading model by visiting https://phala.com/models/z-ai/glm-5.3-flash