Qwen4-Max opens a preview that retires after November 25

128 words · 1 min

Alibaba’s November 11 release gives subscribers a bounded evaluation window for its next flagship model.

Qwen4-Max-Preview supports text, image and video input, tool use and structured output. Access is through Token Plan; ordinary pay-as-you-go access is not supported. [1][2]

The preview starts at 10:00 on November 11 and ends at 24:00 on November 25, both UTC+8. Alibaba says this model will be retired after the preview. Treat it as a comparison opportunity, with a replacement route ready for any experiment that needs to continue. [1]

For developers, the output limit has a practical wrinkle: the advertised maximum of 393,216 tokens is shared by reasoning and the final answer. An output budget sized only for the visible response can therefore truncate a run. The documented context window is one million tokens. [1]

Take it further

Test Qwen4-Max-Preview · Deadline: 26 November 2026Requires a compatible Token Plan subscription. The provider writes the endpoint as November 25 at 24:00 UTC+8, equivalent to November 26 at 00:00 UTC+8. The preview model is scheduled to retire afterward.

Sources & details 3 sources

Discuss in your voice app

Take the reporting, visual explanations and original sources into a conversation. Copy the brief, paste it into your preferred assistant, then start its voice mode.

Open ChatGPT

Paste the brief into your voice app to discuss this reporting.

Reporting sources

Alibaba Cloud — Qwen4-Max-Preview
Original documentation opened through the official release listing; exact preview times, retirement, modalities, limits and budget semantics.
Alibaba Cloud — model releases
Dated November 11 release listing; Token Plan-only preview.
Alibaba Cloud — Token Plan
Subscription and interactive-use scope; no plan purchase or model runtime test performed.

Images & figures

The preview has a fixed end: 25 November at 24:00 UTC+8.
Euangelion · Source · Usage terms