China's AI Models Take Over: A New Weekly Usage Ranking Signals Shift
A fresh set of data from a global multi-model aggregation platform has drawn significant attention. In the latest weekly ranking of AI large language model usage, the top five positions are all held by models developed by Chinese tech firms. This shift points to a movement in technological prowess and suggests that global developer attention is coalescing around new centers of innovation.
Breaking Down the Leaders: Scale and Surge
Topping the chart is Xiaomi's MiMo-V2.5 model, which recorded a massive 10.5 trillion Tokens processed in a single week, marking a 12% increase from the previous period. This scale of usage is a direct indicator of its widespread adoption and practical utility within the developer community.
DeepSeek secured two spots, second and fifth, with a duo of models designed to serve varying needs and complexities. Its flagship Pro version is noted for performance in code generation and complex agent tasks that rivals leading global closed-source alternatives.
The Open-Source Factor: Meteoric Rise of New Entrants
Perhaps more telling is the youth of these top contenders. Each model achieved its prominent ranking shortly after release, demonstrating remarkable market velocity.
Tencent's recently open-sourced model, released just on July 6, delivered the most dramatic surge. It saw a week-over-week usage increase exceeding 999%, making it the fastest-growing model on the list. This "instant-hit" phenomenon underscores the vibrancy of China's open-source AI ecosystem and its capacity to rapidly address developer needs.
Taken together, the data reveals that Chinese AI models are leading not just in absolute usage volume but also in growth momentum and ecosystem vitality. The open-source approach is proving to be a critical pathway for gaining developer trust and expanding technological influence at speed.