AI & Machine Learning

Alibaba Opens Qwen4-Max Preview Through November 25

Qwen4-Max Preview brings a million-token context window to Qwen apps and Token Plan subscribers. Testing ends November 25; a full release date remains unannounced.

By Daniel Mercer Edited by Samantha Reed Published: Updated:
Alibaba Opens Qwen4-Max Preview Through November 25
Alibaba opens Qwen4-Max testing through November 25, with app access and a paid Token Plan API preview. Photo: Qwen

Key Notes

  • Alibaba has opened Qwen4-Max Preview through Qwen apps and Token Plan, with testing scheduled to end at the close of November 25 in Beijing.
  • The model accepts text, images and video, supports a million-token context window and shares its maximum output budget between reasoning and the final answer.
  • API access requires a Token Plan subscription rather than pay-as-you-go billing, and the preview deadline does not establish a full release date or an open-weight launch timetable.

Alibaba has opened a preview of Qwen4-Max, giving users an early look at its next flagship AI model through Qwen’s apps and a subscription service for developers. The test began November 11 and runs through November 25, with no date announced for a full release.

The Qwen team announced early access through QwenWork, Qoder, the Qwen app and QwenStudio. Users can try the model through the free Qwen assistant, and Qwen4-Max-Preview is now listed in the model selector at Qwen Chat. API access follows a separate paid route through Token Plan.

A Million-Token Window for Coding and Agents

QwenCloud’s documentation describes the preview as a multimodal model for reasoning and tasks involving AI agents. It accepts text, images and video, and supports web search, tool calls and structured responses. Its context window holds up to one million tokens.

The maximum output is 393,216 tokens, shared between the model’s reasoning and its final answer. That distinction matters when setting a response budget: reasoning can consume part of the allowance before the user sees the answer. The published specifications establish the model’s limits, but do not by themselves show how reliably it handles demanding work.

API Access Comes With Subscription Rules

For developers, the preview is available through Token Plan rather than ordinary pay-as-you-go API billing. QwenCloud offers individual and team subscriptions that draw from a pool of credits. Supported tools include Qwen Code, Qoder, Cursor and Claude Code, with integration based on compatible API protocols and a dedicated plan key.

Alibaba Cloud’s personal-plan guide lists standard monthly prices of $6, $10, $18 and $43 across four tiers, with temporary introductory discounts on the first two. The service is offered through the Singapore cloud region. Its personal subscription is intended for interactive coding and agent tools, rather than automated scripts, custom application backends or batch processing.

The preview closes at midnight at the end of November 25 in Beijing, equivalent to 16:00 UTC that day. The documentation says the build will remain unchanged during the test and will be retired afterward. That makes the deadline a cutoff for this preview, rather than a confirmed launch date for its successor.

Alibaba previously released open weights for Qwen3.8-Max. The new preview is being offered through hosted services; the current access announcement does not provide a timetable for downloadable Qwen4-Max weights.

Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.

AI & Machine Learning, News