Alibaba's Qwen model downloads exceed 3 billion, tops global open-source AI models
- Alibaba's Qwen open-source model family has surpassed 3 billion cumulative downloads worldwide in the past six months.
- The Qwen series has overtaken models from Meta and Alphabet to become the most-downloaded open-source AI model, according to the report cited by PANews from Bloomberg.
- Alibaba said it has open-sourced more than 460 models in the Qwen series.
- The Qwen ecosystem has generated more than 300,000 derivative models.
- Hugging Face's State of Open Models report said Google models were downloaded about 418 million times in 2026, while Meta models were downloaded about 227 million times.
Alibaba's open-source Qwen model family has crossed 3 billion cumulative downloads globally over the past six months, according to a PANews report citing Bloomberg. The figure puts Qwen ahead of competing open-source models from Meta and Alphabet in current download volume.
Alibaba said the Qwen series now includes more than 460 open-sourced models, and its ecosystem has produced more than 300,000 derivative models. The report said Qwen has become a default workflow for some developers when fine-tuning and deploying models.
The Hugging Face State of Open Models report cited in the article showed Google models were downloaded about 418 million times in 2026 and Meta models about 227 million times. For markets, the news is potentially positive for Alibaba's stock by reinforcing its position in open-source AI, while the broader impact on AI-related equities is neutral to mildly positive.
阿里巴巴通义千问模型下载量突破30亿，居全球开源AI模型首位
据PANews援引彭博报道，阿里巴巴的开源通义千问模型家族在过去六个月内全球累计下载量已突破30亿次。这一下载规模使其在当前开源AI模型中超过Meta和Alphabet的相关模型。
阿里巴巴表示，通义千问系列已开源超过460个模型，生态中还衍生出超过30万个衍生模型。报道指出，通义千问已成为一些开发者在模型微调和部署时的常用选择。
文中援引的Hugging Face《State of Open Models》报告显示，谷歌模型在2026年的下载量约为4.18亿次，Meta模型约为2.27亿次。就市场影响看，这一消息对阿里巴巴股价偏利好，因为它强化了公司在开源AI领域的地位；对更广泛的AI相关板块影响则偏中性到略偏正面。