Dominance in open-source AI

Over 3 Billion Downloads: Alibaba’s Qwen Leaves Meta and Google Behind

Qwen
Facebook
X
LinkedIn
Reddit
WhatsApp
Source: mundissima/Shutterstock.com

Alibaba’s open-source model Qwen records over 3 billion downloads. Chinese AI models leave Meta and Google behind.

According to analyses by developer platform Hugging Face, the open-source model Qwen from Chinese provider Alibaba has surpassed the mark of three billion downloads within six months. This puts the system at the top of global open-source applications and clearly leaves US competitors such as Google (418 million downloads) and Meta (227 million downloads) far behind.

Ad

Through the Qwen family, Alibaba has now published more than 460 individual models, on the basis of which over 300,000 derivative applications have been created worldwide. Hugging Face describes Qwen as “one of the largest foundations of the open AI ecosystem. Qwen has become an integral part of developers’ standard workflow.”

Chinese open-weight systems also surpass US competition when it comes to model sizes: While top models from China reached parameter sizes between 754 billion and 2.78 trillion in 2026, American open-source models remained below the threshold of 130 billion parameters in most months.

Leading role in global token consumption and ecosystem

In addition to Alibaba, other Chinese developers are pushing into the market with high-performance models. Moonshot AI presented Kimi K3, an open-source model with 2.8 trillion parameters, while Zhipu AI released the safety model GLM-5.3 and DeepSeek released the agent framework Harness.

Ad

This adoption is also reflected in ongoing operations. According to evaluations by the OpenRouter service, around 34.25 trillion of the 69 trillion AI tokens consumed worldwide in the first week of August 2026 accounted for Chinese models. This means China has held first place in global token consumption for 15 consecutive weeks. Especially in emerging countries, the multilingual models—Qwen supports 119 languages and dialects—are used in sectors such as education, agriculture, and medicine.

US companies warn against restrictions

The steadily growing dominance of Chinese open-source architectures is leading to debate in the US. In a joint statement, leading US tech companies such as Nvidia, Microsoft, Meta, IBM, and Cisco urged US policymakers to support open model ecosystems and avoid over-regulation.

In addition, the Little Tech Association, an alliance of around 200 Silicon Valley startups, addressed the US government in an open letter to speak out against restrictions on using Chinese open-source models. According to the young companies, a ban would jeopardize their own innovative strength and competitiveness.

(Editorial Team)

Ad

Weitere Artikel