Alibaba's Qwen AI Models Surpass 3 Billion Downloads, Topping Meta and Google

Alibaba's open-weight AI model series Qwen has surpassed 3 billion cumulative downloads globally, overtaking US technology giants Meta and Alphabet as well as Chinese peers to become the most downloaded AI model in the world. In an emailed statement, Alibaba said the Qwen series has open-sourced more than 460 models, with over 300,000 derivative models in its ecosystem. A report released by open-source AI platform Hugging Face on August 14, 2026, showed that Alphabet's Google models had 418 million downloads, while Meta had 227 million.

Open models allow developers to download, customize, and use them as a foundation for building new AI products, making model adoption a key indicator of developers' technology choices. As Chinese developers, including Alibaba, roll out models that are powerful, relatively low-cost, and easy to fine-tune, download counts and derivative model numbers have become metrics of influence in the US-China AI competition. Chinese AI model developers such as Qwen, Moonshot AI, and DeepSeek are striving to achieve frontier-level performance, seeking to narrow the gap with closed-source models from US companies such as OpenAI and Anthropic.

Export controls targeting chips and AI systems - such as the brief ban on overseas users accessing Anthropic's Fable 5 model in the summer of 2026 - have not slowed the progress of Chinese competitors. Hugging Face's report said Qwen's download data makes it one of the largest foundations of the open AI ecosystem, adding that for developers deciding which models to fine-tune and deploy, Qwen has become part of the default workflow.

A broad model family can create a self-reinforcing ecosystem: as more developers adopt these models, derivative versions multiply, attracting new users in turn. Alibaba offers Qwen to enterprise customers in markets such as Southeast Asia and Africa through its cloud platform, further strengthening this cycle and giving it a reach that many competitors lack.