Alibaba launches Qwen4-Max with 1M context and $1.50 input pricing Alibaba's Qwen team launched Qwen4-Max on November 16th, 2026, a flagship model with a 1M-token context window, native vision, and a new mixture-of-experts architecture, priced at $1.50 per million input tokens. Alibaba says Qwen4-Max beats Qwen3.8-Max on its Work360 test while cutting average task cost from $4.43 to $1.67, and positions the model for work, coding, creative tasks and long-running agents. Alibaba launches Qwen4-Max with 1M context and $1.50 input pricing Alibaba says the new flagship beats Qwen3.8-Max on its Work360 test while cutting average task cost from $4.43 to $1.67. By Ryan Merket https://runtimewire.com/author/ryan-merket · Published Alibaba's Qwen team launched Qwen4-Max on November 16th, 2026, positioning the model as its new flagship for work, coding, creative tasks and long-running agents. In a six-post thread on X, Qwen @Alibaba Qwen said Qwen4-Max uses a new mixture-of-experts architecture, handles vision natively and offers a 1M-token context window. The QwenCloud model page describes the same system as a multimodal model for text generation, image and video understanding and deep reasoning, aimed at coding, office productivity, frontend design and complex agentic tasks.…