Alibaba Releases Qwen4-Max, Reports 58% Drop in Task Costs

Alibaba unveiled Qwen4-Max, a flagship model built on a mixture-of-experts architecture with native vision capabilities and a one-million-token context window. The company says it improves on its predecessor in coding, agentic work and professional tasks, and claims it can complete tasks using up to about 50% fewer tokens. Alibaba also reports that average task costs on its Work360 benchmark fell by roughly 60%, attributing the savings to both improved token efficiency and lower API prices; those benchmark figures are company-reported and have not been independently verified in the provided coverage. Bloomberg reports that Alibaba says Qwen4-Max performed near the top of coding benchmarks, ahead of Anthropic’s Claude Fable 5.1 and Moonshot’s Kimi K3.1 Pro. The model is available through Alibaba’s platform and APIs, but Alibaba has not released its weights and says smaller versions are expected soon.




