How can I use Qwen4-Max-Preview?

Qwen4-Max-Preview is available only through Alibaba's Token Plan subscription, and pay-as-you-go API access is not supported. Developers take a dedicated Token Plan API key, point their client at the base URL https://token-plan.maas.qwencloudapi.com and set the model ID to qwen4-max-preview. Token Plan works with Qwen Code, Cline, Claude Code, Cursor, OpenCode and Codex.

How long does the Qwen4-Max preview last?

Qwen's documentation says the Qwen4-Max-Preview window runs from 10:00 on November 11, 2026 to 24:00 on November 25, 2026, UTC+8. Alibaba says the model will not be changed during that period and will be retired when the preview ends, so any setup built on qwen4-max-preview will stop working after November 25.

Are there benchmarks or open weights for Qwen4-Max?

Not yet. At launch Alibaba published no benchmark table, model card, parameter count or license for Qwen4-Max-Preview, and no downloadable weights. The Qwen team's claims of stronger coding, agent work and world knowledge than Qwen3.8 are its own and have not been checked by an outside evaluator.

What are the Qwen4-Max-Preview token limits?

Qwen4-Max-Preview has a 1-million-token context window and returns up to 393,216 tokens, a budget shared by the model's thinking and its final answer. On the Chat Completions API, max_tokens counts both, so Qwen's docs advise leaving room for reasoning when you set it. The model accepts text, images and video and returns text.

Alibaba (Qwen) · 2026-11-11 · major

Qwen4-Max-Preview — Alibaba opens a two-week preview of its next flagship

In short

Qwen4-Max-Preview is the first Qwen4 model anyone can use: a multimodal flagship with a 1M-token context, served only through Alibaba's Token Plan from November 11 to 25. No benchmarks, weights or per-token price yet.

Alibaba (Qwen) · 2026-11-11 · 4 sources checked

Alibaba Cloud

Alibaba's first Qwen4 model goes live as a time-limited preview for coding and agent work.

Quick facts

Maker	Alibaba (Qwen)
Model ID	qwen4-max-preview
Context window	1M tokens
Max output	393,216 tokens (thinking + answer)
Input	Text, images, video
Access	Token Plan only, no pay-as-you-go
Preview window	Nov 11–25, 2026 (UTC+8)

What is it?

Qwen4-Max-Preview is the first model from Alibaba's Qwen4 generation that developers can call. Qwen's docs call it the next-generation flagship multimodal model in the Qwen family, built for coding, office work, front-end design and complex agent tasks. The Qwen team's announcement promises deeper knowledge, richer world understanding and stronger coding and agentic skills than Qwen3.8.

How does it work?

Thinking is on by default and is set with a reasoning_effort parameter (low, medium or xhigh, with xhigh as the default), and thinking shares one output budget with the final answer. The endpoint speaks the OpenAI Chat Completions, OpenAI Responses and Anthropic Messages formats, and supports function calling, web search, built-in tools and JSON Schema output. Calls go through a dedicated Token Plan base URL, so the model can sit behind coding agents such as Qwen Code, Claude Code, Cline or Codex.

Why does it matter?

Developers get an early look at where Qwen's next Max model is heading, inside the coding tools they already use. The catch is in the terms: the preview is frozen, it is retired after November 25, and Alibaba has published no benchmarks, parameter count or license. Treat it as a trial run, not a model to build production work on.

Who is it for?

developers testing coding agents on Alibaba's Token Plan

Frequently asked questions

How can I use Qwen4-Max-Preview?: Qwen4-Max-Preview is available only through Alibaba's Token Plan subscription, and pay-as-you-go API access is not supported. Developers take a dedicated Token Plan API key, point their client at the base URL https://token-plan.maas.qwencloudapi.com and set the model ID to qwen4-max-preview. Token Plan works with Qwen Code, Cline, Claude Code, Cursor, OpenCode and Codex.
How long does the Qwen4-Max preview last?: Qwen's documentation says the Qwen4-Max-Preview window runs from 10:00 on November 11, 2026 to 24:00 on November 25, 2026, UTC+8. Alibaba says the model will not be changed during that period and will be retired when the preview ends, so any setup built on qwen4-max-preview will stop working after November 25.
Are there benchmarks or open weights for Qwen4-Max?: Not yet. At launch Alibaba published no benchmark table, model card, parameter count or license for Qwen4-Max-Preview, and no downloadable weights. The Qwen team's claims of stronger coding, agent work and world knowledge than Qwen3.8 are its own and have not been checked by an outside evaluator.
What are the Qwen4-Max-Preview token limits?: Qwen4-Max-Preview has a 1-million-token context window and returns up to 393,216 tokens, a budget shared by the model's thinking and its final answer. On the Chat Completions API, max_tokens counts both, so Qwen's docs advise leaving room for reasoning when you set it. The model accepts text, images and video and returns text.

Try it

Set model to qwen4-max-preview with base URL https://token-plan.maas.qwencloudapi.com