Alibaba on Monday made its flagship Qwen4-Max model generally available through QwenCloud, following a limited preview that began on November 11. The model uses a mixture-of-experts architecture with 2.4 trillion parameters, accepts images directly and supports a context window of up to one million tokens. It is also available through Qwen Studio, QwenWork and Qoder.

Alibaba says the model uses fewer tokens than Qwen3.8-Max on comparable tasks. Combined with lower API prices, that brought the average task cost on Alibaba’s own Work360 benchmark down from RMB 26.58 to RMB 11.16, a reduction of about 58%. The company also released a plugin that lets agents deliver an initial version of their work and revise it as users give feedback. Alibaba plans to release smaller Qwen4 models as open source. [Qwen]