Alibaba is making bold claims about its Qwen artificial intelligence models, announcing they’ve surpassed 3 billion downloads to become the world’s most-used open-weight AI family. However, independent data from Hugging Face paints a notably different picture, raising questions about how the Chinese tech giant is counting its success.
What Happened
Alibaba announced that its Qwen model family has crossed the 3 billion download threshold, positioning it as the most downloaded open-source AI model globally. The company also claims to have generated more than 300,000 derivative models built on Qwen’s foundation. However, Hugging Face—the primary repository for open-source machine learning models—released data showing significantly lower numbers just one day before Alibaba’s announcement. According to Hugging Face’s metrics, Qwen accumulated approximately 2.045 billion downloads in 2024 alone, with only 151,448 derivatives created by the community.
Key Points
The discrepancy between Alibaba’s figures and Hugging Face’s independent tracking reveals a substantial gap in how success is being measured in the open-source AI space. Alibaba’s derivative count appears nearly double what third-party platforms are registering, suggesting the company may be using different methodologies or including metrics from sources beyond Hugging Face. The timing of Alibaba’s announcement—coming just 24 hours after Hugging Face published contrasting data—intensifies scrutiny around the accuracy of these claims.
This competition reflects the intensifying race among major technology companies to establish dominance in open-source AI development. While Qwen’s popularity remains undeniable, the gap between claimed and verified numbers underscores the importance of independent verification in an industry where marketing claims often outpace documented evidence.
What This Means
For American tech companies and developers, this situation highlights the critical importance of relying on transparent, third-party metrics when evaluating open-source AI models. As enterprises increasingly adopt open-weight models for cost efficiency and customization, accurate download and usage statistics become essential for making informed technology decisions. Alibaba’s ambitious claims shouldn’t overshadow the genuine momentum Qwen has achieved—2 billion downloads through Hugging Face alone represents impressive market penetration. However, stakeholders should approach vendor announcements with appropriate skepticism and cross-reference claims against independent sources like Hugging Face, which provides more auditable metrics.
The broader implication is that the open-source AI landscape remains competitive and dynamic, with multiple players genuinely vying for developer adoption. Rather than relying on inflated figures, the industry benefits when companies demonstrate value through transparent metrics and genuine community engagement.
