Crypto news

Chinese tech giant Alibaba has reached an impressive milestone: the total number of downloads of its open-source AI models from the Qwen family has exceeded 3 billion in just the last six months. To date, the corporation has released more than 460 neural networks into the public domain, which have served as the foundation for 300,000 derivative models created by developers worldwide.

Hugging Face platform data: Qwen's triumph

An analysis of the Hugging Face ecosystem, published in the semi-annual report on the state of open models, paints an even more detailed picture. According to this data, over the first seven months of this year, Qwen models were downloaded 2.05 billion times, and the number of their derivative repositories reached 151,448. For comparison, Google's figure stands at only 82,506. Google model downloads for 2026 are estimated at 418 million, while Meta's are at 227 million.

It is important to emphasize that Hugging Face statistics only account for activity within its own ecosystem, excluding API requests, private deployments, and other distribution channels. The report's authors themselves caution against viewing these numbers as a direct indicator of market share or actual commercial use. Nevertheless, the momentum is impressive: the number of Qwen derivative repositories grows by 180-210 daily. Notably, out of 28,531 GGUF conversions of these models on the platform, only 54 were created by Alibaba itself—the rest is community work.

The secret to success: scaling strategy and licensing policy

Qwen's position is explained by three key factors: regular updates, impressive coverage across all scales—from sub-billion versions to the giant Qwen3.8-Max with 2.4 trillion parameters—and the Apache 2.0 license, which does not restrict modification or commercial use. The breadth of the model lineup has proven decisive. According to Hugging Face data, versions with fewer than 1 billion parameters account for 83% of all downloads in the platform's history, while neural networks over 100 billion account for only 1%.

Laboratories focused exclusively on large LLMs are clearly falling behind. For example, Moonshot AI, which releases almost no models smaller than 70 billion, has gathered only 37 million downloads over the year—roughly 55 times fewer than Qwen. Alibaba also dominates the local deployment segment: its GGUF builds are downloaded 39.6 million times per month, while Gemma's figure is 20.8 million and Llama's is 7.5 million.

A shift in the balance of power in open AI

The report demonstrates a shift in the global balance. In almost every month of this year, the largest open model from China has surpassed all U.S. releases in size. The ceiling for Chinese models ranged from 754 billion to 2.78 trillion parameters, while for American competitors it did not exceed 130 billion in five out of seven months. In China, 59% of models with more than 20 billion parameters are released under Apache 2.0, 22% under MIT, and none of them have restrictions on commercial use. For American developers, the picture is different: 29% under Apache/MIT, 41% under proprietary terms, and 30% without a specified license.

Notably, in the United States, the leaders in the number of new open models were not AI labs but chip manufacturers: AMD and Nvidia each released more than 200 repositories, significantly outpacing Google and Meta.

My analysis: Qwen's success is not mere coincidence but a natural result of a well-thought-out strategy. Alibaba has bet on the democratization of AI, offering the market flexible solutions for any budget and task. This not only strengthens China's position in the AI arms race but also fundamentally changes the rules of the game, forcing Western giants to reconsider their approaches to openness and licensing. While the U.S. focuses on controlling infrastructure, China is capturing the minds and wallets of developers worldwide.