Alibaba has officially released Qwen4, its new-generation base model, with the flagship variant Qwen4-Max now available on the Qianwen AI platform. The release opens API services and Token Plan access to developers and enterprises globally. Qwen4-Max features a new Mixture-of-Experts (MoE) architecture built on Qwen Sparse Attention (QSA), boasting 2.4 trillion total parameters. The model supports native vision capabilities, stable delivery on long-horizon tasks, and million-token context windows.