Alibaba unveiled Qwen3.8-Max-Preview on July 19, 2026, at the World Artificial Intelligence Conference in Shanghai. The model features 2.4 trillion parameters, which Alibaba claims makes it second only to Fable 5.
Developers can access Qwen3.8-Max-Preview through Alibaba’s Token Plan subscriptions and its coding platforms, Qoder and QoderWork. Alibaba did not provide details on the size of the training data or independent benchmark scores for the model.
The launch follows the recent debut of Moonshot AI’s Kimi K3, which has 2.8 trillion parameters. This rapid succession of releases indicates that Chinese AI developers are competing to achieve significant benchmarks.
Alibaba’s previous model, Qwen3.7-Max, was released in May 2026. The company has announced plans to release the model’s weights publicly soon. The Qwen team stated on X that the model is continuously evolving.
Qwen3.8-Max-Preview marks a milestone as the first model in the Qwen series to process images, video, and documents alongside text. However, Alibaba has not disclosed a timeline for the full release or the architecture and context window specifications of the model.
In a related development, Alibaba will serve as a technology partner for Apple Intelligence, which received approval for local release by China’s Cyberspace Administration in July 2026. This partnership is expected to enhance Alibaba’s position in the AI market.
Meanwhile, Moonshot AI is preparing for a Hong Kong initial public offering following the launch of Kimi K3, targeting a valuation near $30 billion. Analysts suggest that the competitive landscape is shifting, with both companies vying to showcase China’s capabilities in AI technology.
