Open-source Qwen models have surpassed the 3 billion download mark: Alibaba's phenomenon and a shift in the global AI landscape
Chinese corporation Alibaba has reached an impressive milestone: downloads of its open-source AI models from the Qwen family have exceeded 3 billion over the past six months. This is a landmark event that fundamentally changes the perception of the balance of power in the open artificial intelligence market.
The scale of the ecosystem is striking: more than 460 Qwen neural networks are publicly available, based on which third-party developers have created about 300,000 derivative models. For comparison, according to the Hugging Face platform, Qwen was downloaded 2.05 billion times over seven months this year, while Google's figure was 418 million and Meta's was 227 million. The number of derivative Qwen repositories reached 151,448 versus 82,506 for Google.
The secret to success: an accessibility strategy
Analysis shows that Qwen's dominance is driven by three key factors. First, regular updates to the model lineup. Second, an incredible breadth of coverage—from compact versions with sub-billion parameter counts to the giant Qwen3.8-Max with 2.4 trillion parameters. It is this factor that proved decisive: models 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%.
Third, the Apache 2.0 license, which imposes no restrictions on modification or commercial use. This favorably distinguishes Qwen from competitors. For example, Moonshot AI, which releases almost no models smaller than 70 billion parameters, has gathered only 37 million downloads over a year—roughly 55 times fewer than Qwen. In the local deployment segment (GGUF builds), Qwen also leads with 39.6 million monthly downloads versus 20.8 million for Gemma and 7.5 million for Llama.
A global shift in the industry
This report clearly demonstrates a shift in the balance of power. Almost every month this year, the largest open-source model from China has surpassed all American releases in size. The difference in licensing policies is also telling: in China, 59% of models with more than 20 billion parameters are released under Apache 2.0, while in the US, 41% of large models are distributed under proprietary terms. Notably, in the US, the leaders in the number of new open models were not AI labs but chip manufacturers—AMD and Nvidia.
My analysis: Qwen's success is not just statistics but a marker of the maturity of China's AI industry. Alibaba bet on the "democratization" of AI, understanding that in the race for developers and real-world adoption, the winner is not the one with the most powerful model but the one offering the most flexible and accessible ecosystem. This is a lesson for Western labs: closedness and gigantism without considering community needs could lead to losing a strategic advantage in the long term.