According to the latest data from the globally renowned AI model aggregation platform OpenRouter, from July 21 to August 11, 2025, the Qwen3 model developed by Alibaba Cloud's Tongyi Qianwen team saw a significant increase in API call volume, with usage surging by 15.4%, demonstrating strong market competitiveness. At the same time, the usage of Anthropic's Claude model decreased by 18.9%, and Google's Gemini model dropped by 6.8%. This trend indicates that Qwen3 is rapidly capturing market share from international AI giants and has become a rising star in the global large model market.
Domestic models are showing remarkable performance, with Kimi and GLM standing out
Aside from Qwen3's strong performance, other domestic large models also performed well. Data shows that the usage of Kimi-K2 increased by 2.8%, while the usage of GLM-4.5 and GLM-4.5-Air models developed by ZAI organization increased by 6%. In comparison, OpenAI's usage only increased by 4.4%. Analysts pointed out that OpenAI's users may be mainly using its official website, applications, or direct API calls, rather than the OpenRouter platform, which could be one of the reasons for its relatively lower growth rate.
Qwen3-Coder becomes the market leader
As the flagship model in the Qwen3 series, Qwen3-Coder quickly became a market focus due to its excellent performance in code generation, tool calling, and long-context reasoning. According to OpenRouter data, Qwen3-Coder ranked first with nearly 50 billion tokens of call volume. The Tongyi Qianwen series models occupied the top three positions and secured five spots among the top ten models, with total call volume exceeding 200 billion tokens. Industry insiders stated that Qwen3-Coder has surpassed top models like GPT-4.1 and Claude4 in multiple agent capability evaluations, demonstrating its strong competitiveness in the programming field.
The "Android Moment" of Chinese AI is coming
Some industry insiders refer to July 2025 as the "Android Moment" of Chinese large models, indicating comprehensive breakthroughs in scale, performance, and cost for Chinese AI models. Models such as Qwen3, Kimi-K2, and GLM-4.5 have rapidly spread through open-source formats, exerting a profound impact on the global AI market landscape with their high performance and low cost. Meanwhile, the market share of Claude and Gemini continues to decline, reflecting the rising momentum of Chinese AI companies in global competition.
Future Outlook: The Global Journey of Chinese AI
With the rapid growth of domestic models like Qwen3, the Chinese AI industry is showing strong competitiveness on a global scale. OpenRouter data shows that Tongyi Qianwen has a market share of 10.4%, surpassing OpenAI's 4.7% and ranking fourth globally. In the future, whether Qwen3 and its series models can further challenge the position of international giants remains to be closely watched.
Summary
The rapid rise of Qwen3 marks a strong breakthrough of the Chinese AI industry in the global market. With the decline of Claude and Gemini, domestic large models are reshaping the market landscape through technological innovation and cost advantages. In the future, this trend is likely to intensify, and AIbase will continue to keep you updated on the latest developments in this field.
