BlockBeats News, August 26, Alibaba's Qianwen released the Qwen3.8-Flash model. According to the introduction, this is a multi-modal MoE model and an early preview version of the Qwen4 architecture. The production version of Qwen3.8-Flash will soon be available through the Qwen Cloud API, priced at only $0.16 per million input tokens and $0.47 per million output tokens. The model has 1.25 trillion parameters + 51 billion N-gram embedding parameters, but only activates 60 billion parameters per token, achieving high cost-effectiveness.