Amid the rapid iteration of global AI technology, the Chinese large model market is showing extraordinary explosive power and application resilience. According to the latest industry estimates released by OpenRouter, during the week of May 4 to May 10, the total number of tokens used for global AI large models reached an astonishing 25.7 trillion tokens, a 7.5% increase from the previous week, showing a strong upward trend for three consecutive weeks.

In this global competition for computing power and applications, the performance of Chinese large models stands out particularly. Data shows that last week, the weekly token usage of Chinese AI large models reached 7.941 trillion tokens. Although the scale remained basically the same as the previous week, its massive base has established an absolute market advantage.

By comparison, the weekly token usage of American AI large models during the same period was 3.76 trillion tokens. Although the US market achieved a significant 14.41% month-on-month growth during the week, there is still a significant gap between the total volume and China's. From the data comparison, the weekly token usage of Chinese large models has already reached 2.11 times that of the United States.

This is not just a simple numerical lead, but also marks that Chinese large models have continuously exceeded those in the United States in terms of usage scale for two consecutive weeks and firmly hold the top position globally. This sustained leading trend reflects that the application scenarios of domestic large models are currently in a highly active period, and developers and companies are increasingly relying on and utilizing local AI capabilities.