Skip to content

Alibaba previews Qwen4-Max, then retires it in two weeks

Qwen4-Max-Preview ships in Alibaba Cloud Model Studio with a 1M-token context and OpenAI- and Anthropic-compatible APIs, but it is cloud-only, Token Plan-only, and switches off on 25 November.

Source: Alibaba Cloud

Alibaba put the first preview of its next flagship model, Qwen4-Max, into Alibaba Cloud Model Studio on 11 November. There are no open weights, and access runs out on 25 November, when the preview id is switched off for good.

The model takes a 1 million token context window and caps output at 393,216 tokens, shared between its thinking and its final answer. Thinking is on by default, controlled through a reasoning_effort parameter set to low, medium or xhigh. The API is compatible with OpenAI’s Chat Completions and Responses formats and with Anthropic’s Messages API, though Alibaba redefines max_tokens to cover thinking plus answer combined, rather than just the answer.

Why it matters: an OpenAI- and Anthropic-shaped API makes swapping this into an existing integration for a benchmark run close to a drop-in change, and a 1M token context at the top tier is worth testing against whatever you use now for long-document or agentic work while the preview is live.

The caveat: there is no pay-as-you-go access. You need a Token Plan subscription and a dedicated API key, and after 25 November the model id simply stops responding, with no stated replacement yet.

reasoning_effort: "low" | "medium" | "xhigh"
Share

Related reading

More from The Wire

All briefs