What happened?
The open-weight Qwen model family from Alibaba Group Holding surpassed 3 billion downloads globally over the past six months. According to a report by Bloomberg reporter Saritha Rai, this figure has positioned Alibaba as the world's most preferred provider in the open-source AI model space.
According to Hugging Face's "State of Open Models: Summer 2026" report, during the same period Google reached 418 million downloads and Meta Platforms reached 227 million. This data shows Alibaba has created a significant gap compared to its rivals.
Why does it matter?
According to Hugging Face's findings, there are 151,448 Qwen-based derivative models. This number is 2.6 times Meta's total model ecosystem and 4.7 times the number of Llama-based repositories. This indicates that Qwen has become not just a downloaded model, but a foundation model that developers are building upon.
Jun Song, a user who shared the topic on social media, noted that the release of Qwen 3.5 sparked a massive surge of interest in local AI models. Another assessment stated that the Qwen 3.8 27B version achieved the anticipated major leap, surpassing companies like Meta and Alphabet as well as local competitors.
What we know
- The Qwen model family received over 3 billion global downloads in six months.
- Google reached 418 million downloads in 2026, while Meta reached 227 million.
- Hugging Face identified 151,448 Qwen-based derivative models.
- This figure is 2.6 times Meta's total footprint and 4.7 times the size of Llama repositories.
Comparison: Download Figures
| Company | Downloads (6 months / 2026) |
|---|---|
| Alibaba (Qwen) | 3 billion+ |
| 418 million | |
| Meta | 227 million |
What's next?
Hugging Face's report serves as a regular observation revealing the scale of the open-weight model ecosystem and its impact on the developer community. Whether Alibaba can sustain this momentum will become clearer as new Qwen versions are released and rival companies pursue their own open-source strategies.