Alibaba's Qwen AI Models Now Account for Over Half of All Global Open-Source Downloads

Qwen Dominates the Open-Source AI Landscape

Alibaba Cloud has solidified its position as the world's most widely adopted open-source AI model provider. A report released on the tenth of April by US-based tracking organisation Interconnects AI reveals that the Qwen model series now accounts for more than fifty percent of all global open-source model downloads, with cumulative downloads approaching nine hundred and forty-two million.

In February alone, Qwen generated one hundred and fifty-three point six million downloads, more than double the combined total from the next eight largest providers, including Meta, DeepSeek, and OpenAI.

Aggressive Pricing and Rapid Releases

Alibaba has positioned speed and affordability as core differentiators. The company launched Qwen three point five Plus in February at an API price roughly one-eighteenth the cost of Google's Gemini three Pro. In early April, Alibaba followed up with Qwen three point six Plus, a closed-source model tuned for agentic coding with a one-million-token context window, marking a strategic shift toward monetisation.

From Open-Source Champion to Enterprise Monetisation

The milestone comes amid a significant strategic pivot. Alibaba is increasingly moving its most capable models behind closed-source paywalls, launching enterprise platforms like Wukong for agentic AI deployment and Accio Work for business productivity. The shift is driven by intensifying competition from ByteDance in China's cloud market.

Investing in the Next Frontier

Alibaba also led a roughly two billion yuan investment in ShengShu Technology, the startup behind the AI video generation tool Vidu, signalling a push into world models that combine visual understanding with physics-based reasoning. Meanwhile, the company has activated a new data centre equipped with ten thousand in-house Zhenwu AI chips, reducing its dependence on Western semiconductor technology.