Qwen4-Max Preview Opens With 1M Context
Alibaba opened a preview of Qwen4-Max, the first Qwen 4 model, on November 11, 2026. It accepts text, images and video with a 1M-token context, is available only through QwenCloud's Token Plan from $6 a month, and will be retired after 24:00 on November 25 (UTC+8).
Alibaba opened a two-week preview of Qwen4-Max on November 11, 2026, the first model in its next-generation Qwen 4 family. The model reads text, images and video, accepts up to 1 million tokens of context, and is available only through QwenCloud's Token Plan subscription until 24:00 on November 25 (UTC+8), after which it will be retired, according to QwenCloud's developer documentation.
Key takeaways
- Qwen4-Max-Preview runs from 10:00 on Nov. 11 to 24:00 on Nov. 25, 2026 (UTC+8); Alibaba says the model will not change during that window and will be retired afterward.
- It offers a 1M-token context window and up to 393,216 output tokens, a budget shared by reasoning and the final answer.
- Access is Token Plan only, with no pay-as-you-go option; individual plans start at $6 a month.
Qwen4-Max-Preview specs: context, output and modes
QwenCloud describes the model, ID qwen4-max-preview, as "the next-generation flagship multimodal model in the Qwen family," aimed at coding, office productivity, front-end design and complex agentic tasks. An alias, qwen4.0-max-preview, also works.
Thinking is on by default. Developers can set reasoning effort to low, medium or xhigh, with xhigh as the default. The model supports function calling, web search, built-in tools and structured output in JSON Object and JSON Schema formats.
It is compatible with OpenAI's Chat Completions and Responses APIs and Anthropic's Messages API, which lets existing agent tools point at it by swapping a base URL and key.
How to access Qwen4-Max: Token Plan pricing
Pay-as-you-go API access is not supported. Token Plan is sold only in QwenCloud's Singapore region, uses its own endpoint and keys prefixed "sk-sp-," and works with tools such as Claude Code, Codex, Cursor, Qwen Code and OpenClaw, according to the Token Plan overview.
| Tier | Monthly price | Monthly Credits |
|---|---|---|
| Lite | $6 ($3 new-user offer) | 20,000 |
| Essential | $10 ($5 new-user offer) | 45,000 |
| Standard | $18 | 80,000 |
| Pro | $43 | 200,000 |
During the preview, Qwen4-Max calls consume Credits at as little as 20% of the standard rate, and calls between 22:00 and 08:00 (UTC+8) drop to 8%, the plan page says. The new-user prices run until November 30. TechNode reported that the model is also available for early use in the Qwen app, Qoder, QwenWork and QwenStudio.
A max_tokens change that can break code
A QwenCloud compatibility notice dated November 2 says that, starting with qwen4-max-preview, max_tokens in the OpenAI-style Chat Completions API equals max_completion_tokens and caps reasoning and answer tokens combined. A limit sized only for the answer may now cut responses short. Older Qwen models keep their previous behavior.
Context: what Alibaba has not published
Alibaba said at its 2026 Apsara Conference that Qwen 4 was in training and that its Qwen 4.5 and Qwen 5 series are projected to scale to 5 trillion to 10 trillion parameters. The Qwen4-Max documentation lists no parameter count, benchmark scores or per-token price.
Independent tracker ModelFit reported that no Qwen4 weights appeared in the Qwen organization on Hugging Face as of November 11, so the preview cannot be run locally.
What's next
Alibaba has not announced a general-availability date, final pricing or an open-weight release for Qwen 4. Developers testing the preview have until November 25 before the model is retired.
FAQ
How can I use Qwen4-Max-Preview?
Subscribe to a QwenCloud Token Plan, generate a plan key, use the Token Plan base URL and set the model ID to qwen4-max-preview. Pay-as-you-go keys do not work.
Can I download Qwen4-Max weights?
No. Alibaba has not released weights, and the preview runs only on its cloud.
When does the preview end?
At 24:00 on November 25, 2026 (UTC+8). The model will be retired after that.
References
- Latest model: Qwen4-Max-Preview - QwenCloud — QwenCloud
- Token Plan Individual - QwenCloud — QwenCloud
- API compatibility upgrades - QwenCloud — QwenCloud (Nov. 2, 2026)
- Alibaba Cloud's 2026 Apsara Conference: Full-Stack AI Roadmap Along with Global Market Expansion Plan — Alibaba Cloud
- Alibaba opens Qwen4-Max preview with a million-token context window — TechNode (Nov. 11, 2026)
- Qwen4-Max Preview: 1M Context, Cloud Only, No Weights — ModelFit (Nov. 11, 2026)
Editor responsible for AI Newsway. Writes about AI, developer tools, cloud and productivity technology.
Comments 0
Comments are currently disabled.