According to the latest monitoring data, global AI large models experienced explosive growth last week, with a total of 36.1 trillion Tokens used, marking seven consecutive weeks of growth. In this wave of artificial intelligence enthusiasm, Chinese AI large models have stood out particularly.
Data shows that the weekly usage of Chinese AI large models reached 14.19 trillion Tokens last week, increasing by 27.49% compared to the previous week. Meanwhile, the weekly usage of US large models declined. China has maintained a leading position in usage for six consecutive weeks, firmly holding the top spot globally.
Domestic Large Models Dominate the Top Four
In the global ranking of large model usages, Chinese domestic models have occupied the top four positions. Among them, DeepSeek-V4-Flash performed exceptionally strongly, maintaining the top position for three consecutive weeks.
The second place is held by Tencent Hy3 preview, which has also remained in the second position for three consecutive weeks. MiniMaxM3, a newly released model, has become the biggest surprise, entering the global top three in its first week due to its strong programming capabilities and ultra-long context window.
Market Structure Facing Reconfiguration
Under the impact of the comprehensive rise of domestic large models, the US Claude series models, which previously held an important position in the developer ecosystem, have shown a significant decline in shipments. Their main models have not only fallen out of the top five but also experienced a substantial decrease in weekly usage.
Looking at the overall shipment performance of manufacturers, Chinese domestic companies such as DeepSeek and MiniMax have demonstrated extraordinary dominance. Chinese large models, driven by technological innovation and high cost-effectiveness, are accelerating the reshaping of the global artificial intelligence market competition landscape.
