Chinese AI Models Dominate OpenRouter's Top 5 in Weekly Usage
A recent weekly ranking of AI large model usage on the global multi-model aggregation platform OpenRouter shows that the top five positions are all occupied by models developed by Chinese companies. Leading the list is Xiaomi's MiMo-V2.5, which recorded 10.5 trillion tokens in weekly calls, a 12% increase week-over-week. DeepSeek secured both the second and fifth spots with two models designed to meet different development needs. Their flagship Pro version is noted for its performance in code generation and complex agent tasks, rivaling top overseas closed-source models. A common characteristic of these models is their rapid ascent to the top rankings shortly after their release. Tencent's Hunyuan 3, which officially became open-source on July 6th, claimed the third position with an impressive week-over-week growth exceeding 999%, making it the fastest-growing model on the list.
The rapid rise of Chinese large language models on global aggregation platforms like OpenRouter, particularly those with recent open-source releases, highlights a significant shift in the AI development landscape. This trend suggests that Chinese tech firms are increasingly competitive in producing powerful and accessible AI models, challenging established global players. The swift adoption and high usage figures indicate strong developer interest, potentially driven by performance, cost-effectiveness, or specific capabilities tailored to emerging market needs. As these models mature and gain wider adoption, their influence on global AI development, research, and application ecosystems is likely to grow, fostering greater diversity and competition within the field.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.