Alibaba Launches Qwen3.8-Max Model

On August 3, Alibaba officially released the Qwen3.8-Max AI model, with total parameters expanded to 2.4 trillion. The model weights are scheduled to be open-sourced next week. According to reports, Qwen3.8-Max delivers comprehensive improvements in programming, real-world office tasks, long-horizon tasks, and multimodal agent capabilities. It is capable of end-to-end completion of complex, open-ended goals with minimal human intervention, delivering trustworthy results. Qwen3.8's API pricing for domestic users are as follows: input costs 12 RMB per million tokens, output costs 36 RMB per million tokens, and implicit cache hits cost 1.5 RMB per million tokens.