On July 19, Alibaba's Qwen team previewed Qwen3.8-Max-Preview, a new flagship multimodal model with 2.4 trillion parameters. The announcement was made during the World AI Conference in Shanghai, arriving two days after Moonshot AI released its Kimi K3 model.
- The model processes text, images, video, and documents, marking it as Alibaba's first multimodal model exceeding 1 trillion parameters.
- It uses a sparse Mixture-of-Experts architecture and is compatible with OpenAI and Anthropic protocols.
- The preview is available via Alibaba's Token Plan subscription at 10% of standard pricing.
- Alibaba claims the model is second only to Fable 5 in performance, though no benchmark table or license has been published.
- The total parameter count is unverified, and the active parameter count per token remains undisclosed.
The preview allows developers to test the model's capabilities in coding, full-stack development, data analysis, and office workflows before any official open-weight release.