Alibaba releases Qwen4-Max AI model with reduced API pricing
Alibaba has made its flagship Qwen4-Max model generally available to the public. The update introduces a million-token context window and significantly lower operational costs for enterprise users.

Alibaba has moved its flagship Qwen4-Max AI model into general availability on its QwenCloud platform following a brief preview period. The model utilizes a mixture-of-experts architecture with 2.4 trillion parameters and features native image processing capabilities alongside a context window capacity of one million tokens.
The release emphasizes cost reduction for corporate API consumers. According to Alibaba, the model demonstrates improved token efficiency compared to the previous Qwen3.8-Max iteration. When coupled with lower API unit pricing, the company reports that average task costs on its internal Work360 benchmark have decreased by approximately 58 percent.
In addition to the model release, Alibaba introduced a plugin designed to enable iterative refinement for AI agents based on direct user input. The company also confirmed plans to transition smaller versions of the Qwen4 series to an open-source distribution model in the future.