News

Alibaba launches Qwen4-Max multimodal model across six regions

Alibaba Cloud has released Qwen4-Max, a multimodal model that accepts text, images and video, through Model Studio deployments in six regions.

D
Nov 15, 2026 · 1 min read

Alibaba Cloud released Qwen4-Max on November 16 and made the multimodal model available through Model Studio in six documented regions. Alibaba calls it the Qwen family’s ‘next-generation flagship,’ but that label is the company’s positioning, not an independent comparison with competing models.

According to the Qwen4-Max documentation, the invocation ID is qwen4-max. The model accepts text, image and video inputs and returns text. The non-preview release follows a separate, time-limited Qwen4-Max-Preview program.

Alibaba lists deployments in China (Beijing), Singapore, Germany (Frankfurt), the US (Virginia), Japan (Tokyo) and Hong Kong. Beijing has a Chinese-mainland scope, Singapore has an international scope, and Frankfurt, Virginia, Tokyo and Hong Kong have a global scope.

The company documents a 1,000,000-token context length, with maximum input and output limits that vary by thinking mode. Model Studio’s listed capabilities also include streaming, deep thinking, parallel tool calling, web search, function calling and structured output.

According to the documentation, Qwen4-Max supports OpenAI Chat Completions, Responses, Anthropic Messages and DashScope interfaces. Those references establish API compatibility; they do not show that Qwen4-Max matches or exceeds models sold by OpenAI or Anthropic.

A separate Alibaba Cloud product changelog records Qwen4-Max as available on November 16. Its separate November 11 entry for Qwen4-Max-Preview distinguishes the earlier preview from the production model documented under the qwen4-max ID.

More news