Chinese AI Models Have Swept Global Token Rankings
Chinese artificial intelligence models captured all of the top five spots in weekly global token usage rankings.
Updated on Oct. 12, 2026 in Artificial Intelligence
Live Poll
Should your country limit reliance on foreign-developed AI models for national digital infrastructure?
Chinese AI models processed a dominant 86.74 trillion tokens globally during the week of October 5 to 11, 2026. This surge allowed these platforms to secure the top five positions in usage rankings, outpacing their counterparts from the United States.
Why it matters
The shift highlights a major trend as global developers increasingly turn to low-cost, open-weight models from Chinese companies for their AI projects. This data marks a significant change from February 2026, when Chinese models accounted for only 6 to 13 percent of total usage.
DeepSeek V4.1 Flash led the market by processing 41 trillion tokens, while GLM 5.3 Flash and MiMo-V2.6-Flash processed 12.2 trillion and 11.5 trillion tokens, respectively. These models combined to facilitate a significant portion of the 172 trillion tokens processed globally that week.
The players
DeepSeek
DeepSeek is a Chinese artificial intelligence company that developed the V4.1 Flash model.
Z.ai
Z.ai is an AI developer responsible for the GLM 5.3 Flash language model.
Xiaomi
Xiaomi is a multinational technology corporation that produced the MiMo-V2.6-Flash AI model.
StepFun
StepFun is an artificial intelligence firm that created the Step 5 Preview model.
Tencent
Tencent is a major Chinese technology conglomerate that released the Hy4 preview model.
The details
OpenRouter aggregated these figures, revealing that while global usage grew by 3.61 percent week-over-week, Chinese models saw a much steeper adoption curve. Developers are prioritizing these specific Chinese models for their technical capabilities and cost-effectiveness compared to established alternatives.
Timeline
February 2026: Chinese models accounted for 6-13 percent of total usage.
Mid-August 2026: DeepSeek increased API prices for its flagship models.
October 5-11, 2026: Chinese models occupied the top five spots for weekly token usage.
The Tech Race
This dominance represents a major pivot in the global AI landscape, as low-cost, open-weight models rapidly displace the legacy dominance of premium proprietary systems. It mirrors a broader evolution in tech where accessibility and deployment scale have become more critical to market share than traditional brand recognition.
Developers and businesses globally can now access higher-performing AI tools at significantly lower price points than previously available. This accessibility allows for the rapid integration of advanced AI capabilities into a wider array of consumer and enterprise software applications.
The takeaway
The rapid ascent of Chinese AI models suggests that cost-efficiency is currently the primary driver for global developer adoption. Users should evaluate both the performance metrics and the evolving cost structures of these models when selecting tools for their development stacks.
Further reading
For more on the evolving metrics of language model adoption, explore our Artificial Intelligence section.
Live Poll
Should your country limit reliance on foreign-developed AI models for national digital infrastructure?
