Xiaomi Group recently released data showing that the Xiaomi MiMo-V2.5 large model has risen to the top in the global AI model call volume ranking. The weekly call volume reached 10.5 trillion Tokens during the week of July 20–26, representing a 12% increase compared to the previous week, making it one of the most widely used AI large models globally.

Xu Jieyun, Special Assistant to the Chairman of Xiaomi Group and Deputy Director of the Strategic Marketing Department, revealed that the top five positions in the global AI model call volume ranking were all occupied by Chinese models last week, indicating that the penetration speed of Chinese AI large models in practical applications and developer ecosystems continues to rise. Among them, Xiaomi MiMo-V2.5 achieved a significant ranking improvement due to its increased call volume, becoming the biggest change in this ranking.

62a268b4ly1ifidj1t8bqj20mi0u00wb.jpg

Token call volume is typically considered one of the important indicators for measuring the actual usage scale of AI models, reflecting the model's activity in application scenarios, developer access, and enterprise services. The fact that MiMo-V2.5 reached 10.5 trillion Token weekly call volume indicates that the demand for its use in smart applications, development interfaces, and related business scenarios is further expanding.

In recent years, as the competition among large models has shifted from parameter scale to reasoning capabilities, cost efficiency, and ecological applications, Chinese AI companies have continuously promoted the commercialization of models. Domestic large models, including those from Xiaomi, Alibaba, ByteDance, and DeepSeek, are accelerating their entry into the global AI infrastructure competition by opening up interfaces, reducing call costs, and expanding application scenarios.

The recent top ranking of Xiaomi MiMo in the global call volume list also reflects that AI model competition is shifting from technical indicator comparisons to real user scale and ecological influence competition. In the future, model stability, developer ecosystems, and industry application capabilities will become important factors in evaluating the long-term competitiveness of large models.