Qwen's Expansion: 3 Billion Downloads and a Shift in Leadership in Open AI
Chinese tech giant Alibaba has reached an impressive milestone: the cumulative number of downloads of their open AI models from the Qwen family over the past six months has exceeded 3 billion. This is not just a number, but a marker of a tectonic shift in the industry, where open models from China are beginning to dominate their Western counterparts in popularity and scale of distribution.
The corporation has made more than 460 neural networks publicly available, which have served as the foundation for developers worldwide to create approximately 300,000 derivative models. Such an ecosystem creates a powerful network effect, strengthening Qwen's position as the de facto standard for fine-tuning and customization.
Hugging Face Data: China vs. the West
Analytics from the Hugging Face platform, published ahead of this announcement, paint an even more detailed picture. Over seven months of the current year, Qwen downloads totaled 2.05 billion, and the number of derivative repositories reached 151,448. For comparison, Google's figure stood at 82,506. Google's cumulative model downloads for 2026 are estimated at 418 million, and Meta's at 227 million, which is significantly lower than Alibaba's figures.
It is important to understand that Hugging Face statistics only account for activity within their own ecosystem, excluding API requests and private corporate deployments. Despite this, the gap is so large that it clearly indicates a shift in priorities within the developer community.
Success Factors: From Licenses to Scale
Qwen's success is no accident. I see three key factors that drove this breakthrough. First, it is an impeccable licensing policy: Apache 2.0 imposes no restrictions on commercial use and modification, which is critical for businesses. Second, it is the incredible breadth of the model lineup—from compact versions with sub-billion parameters to the flagship Qwen3.8-Max with 2.4 trillion parameters. It is the small models that generate the bulk of traffic: they account for 83% of all downloads in the platform's history, while giants over 100 billion account for only 1%.
Third, it is the regularity of updates. Alibaba is not standing still, releasing new versions with enviable frequency, which forces the community to constantly return to their ecosystem. In the local deployment segment (GGUF builds), Qwen also leads with 39.6 million monthly downloads, surpassing Gemma (20.8 million) and Llama (7.5 million).
Balance of Power and the Future of the Industry
The report confirms a shift in the balance: in five of the seven months of the current year, the largest open model from China surpassed all American releases in size. The ceiling for Chinese models ranged from 754 billion to 2.78 trillion parameters, while for US competitors it did not exceed 130 billion in most cases. Notably, in the US, the main generators of open models were not AI labs but chip manufacturers—AMD and Nvidia, which each released more than 200 repositories.
From my point of view, this is only the beginning. Control over infrastructure and chips remains with the US, but in the battle for developers' minds and real AI adoption, China has already won a convincing victory. The openness and accessibility of Qwen create a powerful foundation for the future dominance of Chinese technology on a global scale, and Western companies will have to rethink their strategy to avoid losing this market entirely.