China Extends Lead in Global AI Model Usage, Tencent Hunyuan New Model Tops List
OpenRouter, a third-party data platform, reports that global AI large models consumed 26.9 trillion tokens last week, marking four consecutive weeks of strong growth. Amid the global race for AI computing power and applications, Chinese AI large models have distinguished themselves, sustaining a high weekly usage volume.
While weekly token usage of American AI large models rose sharply from the prior week, Chinese models' weekly usage volume now stands at 1.81 times that of the US. This marks the third consecutive week Chinese large models have outpaced American ones, securing the global lead.

Tencent's Model Takes the Lead, Price Cuts Highlight Inference Cost Advantage
Among the top three models by usage volume, the first two are Chinese AI large models. Tencent's Hy3 preview model recorded 2.66 trillion tokens in weekly usage, securing the top spot and maintaining strong user retention after its free trial period.
DeepSeek, a rising domestic star, also performed strongly, with three of its models remaining on the list and usage volumes climbing. DeepSeek-V4-Flash reached 2.06 trillion tokens per week, ranking second globally. Price reductions are helping Chinese large models draw more international users.
Anonymous Agent Model Surprises, High-Performance Base Models Gain Popularity
Beyond the direct competition between Chinese and US giants, an anonymous model named 'Owl Alpha' appeared on last week's list, quickly rising to eighth place. Its weekly token usage doubled, making it one of the most talked-about tech newcomers.
Industry platforms reveal that Owl Alpha is a high-performance base model built for Agent (intelligent entity) workflows. It supports complex tool calls, ultra-long context of millions of tokens, and excels at code generation and executing complex instructions.
Related article
Qwen3.5-Max Preview Tops LMArena as China's AI Models Reach Global Front Rank
On March 20, the latest rankings from the authoritative large model blind test platform LMArena revealed that Alibaba's newest flagship preview model in the Qwen series, Qwen3.5-Max-Preview, delivered exceptional performance. Scoring 1464 points, it
Marc Benioff-backed startup aims to solve AI deployment challenges
Big enterprises face such steep challenges in deploying reliable AI tools that entire teams of forward-deployed engineers—specialists embedded within client companies to operationalize AI systems—are emerging to bridge the gap.“Paradoxically, AI driv
AliVideo Unveils Wan3.0, Delivering Ultra-Long Video Generation and Multi-Dimensional Consistency
On August 24, Alibaba unveiled Wan3.0, its latest video generation large model. This release represents a major leap forward for Alibaba’s video technology, delivering significant improvements in duration, universal creation, reference handling, and
Related Special Topic Recommendations
Comments (0)
0/500
OpenRouter, a third-party data platform, reports that global AI large models consumed 26.9 trillion tokens last week, marking four consecutive weeks of strong growth. Amid the global race for AI computing power and applications, Chinese AI large models have distinguished themselves, sustaining a high weekly usage volume.
While weekly token usage of American AI large models rose sharply from the prior week, Chinese models' weekly usage volume now stands at 1.81 times that of the US. This marks the third consecutive week Chinese large models have outpaced American ones, securing the global lead.

Tencent's Model Takes the Lead, Price Cuts Highlight Inference Cost Advantage
Among the top three models by usage volume, the first two are Chinese AI large models. Tencent's Hy3 preview model recorded 2.66 trillion tokens in weekly usage, securing the top spot and maintaining strong user retention after its free trial period.
DeepSeek, a rising domestic star, also performed strongly, with three of its models remaining on the list and usage volumes climbing. DeepSeek-V4-Flash reached 2.06 trillion tokens per week, ranking second globally. Price reductions are helping Chinese large models draw more international users.
Anonymous Agent Model Surprises, High-Performance Base Models Gain Popularity
Beyond the direct competition between Chinese and US giants, an anonymous model named 'Owl Alpha' appeared on last week's list, quickly rising to eighth place. Its weekly token usage doubled, making it one of the most talked-about tech newcomers.
Industry platforms reveal that Owl Alpha is a high-performance base model built for Agent (intelligent entity) workflows. It supports complex tool calls, ultra-long context of millions of tokens, and excels at code generation and executing complex instructions.
Qwen3.5-Max Preview Tops LMArena as China's AI Models Reach Global Front Rank
On March 20, the latest rankings from the authoritative large model blind test platform LMArena revealed that Alibaba's newest flagship preview model in the Qwen series, Qwen3.5-Max-Preview, delivered exceptional performance. Scoring 1464 points, it
Marc Benioff-backed startup aims to solve AI deployment challenges
Big enterprises face such steep challenges in deploying reliable AI tools that entire teams of forward-deployed engineers—specialists embedded within client companies to operationalize AI systems—are emerging to bridge the gap.“Paradoxically, AI driv
AliVideo Unveils Wan3.0, Delivering Ultra-Long Video Generation and Multi-Dimensional Consistency
On August 24, Alibaba unveiled Wan3.0, its latest video generation large model. This release represents a major leap forward for Alibaba’s video technology, delivering significant improvements in duration, universal creation, reference handling, and





Home






