Alibaba Unveils Qwen3.8-Max, Its Largest AI Model Yet

The new model features 2.4 trillion parameters, placing it close to domestic rival Moonshot AI’s recently launched Kimi K3, which has 2.8 trillion parameters. While parameter count is widely regarded as an indicator of an AI model’s scale and computational power, it does not necessarily determine overall performance. Chinese AI firms increasingly highlight these figures to attract developers, particularly as most of their models are released with open weights, allowing researchers and businesses to download, modify, and deploy them. In contrast, companies such as OpenAI, Anthropic, and Google do not disclose parameter counts for their proprietary models.

Qwen3.8-Max made a strong debut on the Arena.AI model comparison platform, becoming the highest-ranked Chinese text model. Globally, it trails only Anthropic’s Claude Fable 5 and three Opus variants. In the platform’s vision-model rankings, which assess AI systems capable of analyzing images and visual content, the model secured second place worldwide behind a Claude Fable 5 variant.

The multimodal model can process text, images, and video, while supporting a context window of up to one million tokens. This enables it to analyze extensive datasets, including lengthy legal documents, software codebases, and hundreds of pages of text in a single session.

Alibaba said Qwen3.8-Max employs a mixture-of-experts architecture, activating only 95 billion parameters for each task instead of the model’s full 2.4 trillion. The approach is designed to lower computing costs and improve response speed. The company also claimed the model successfully completed a software engineering project in just 16 days.

Qwen3.8-Max is scheduled for release next week through Alibaba Cloud’s Model Studio platform. – ERMD

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top