Alibaba has launched a preview of Qwen4-Max, its next-generation flagship AI model, with a context window of up to 1 million tokens. The model can process text, images, and videos. It is designed to handle complex coding tasks, long documents, research, and AI agent workflows.
Announced on November 11, 2026, Qwen4-Max is the first preview release in Alibaba’s upcoming Qwen4 model family. The company is offering early access through its Token Plan subscription and several AI platforms, including the Qwen app and Qoder.
The launch adds a new flagship model to Alibaba’s AI lineup as competition grows among companies developing advanced AI models. Major areas of competition include longer context windows, better reasoning, and stronger coding performance.
Alibaba Qwen4-Max Supports 1 Million-Token Context Window
Alibaba’s Qwen4-Max AI model supports a context window of up to 1 million tokens, allowing it to process large volumes of information in a single conversation. This makes the model suitable for tasks that require analyzing lengthy documents, large codebases, and multiple sources of information.
The larger context window could help developers review entire software projects, summarize long reports, and complete complex tasks without splitting information into smaller sections. Qwen4-Max also supports image and video inputs, allowing users to analyze visual content alongside text. This expands its capabilities beyond text-based AI interactions.
According to Alibaba Cloud’s documentation, Qwen4-Max includes deep reasoning, function calling, web search, and structured output capabilities. The model also supports OpenAI’s Chat Completions and Responses APIs, along with Anthropic’s Messages API. This allows developers to integrate Qwen4-Max into existing applications with fewer compatibility changes.
ALSO READ: Alibaba AI Models Compared (2026): Qwen 2.5 & More
Alibaba Qwen4-Max Preview Availability, Access, and API Details
Alibaba has released Qwen4-Max Preview through its Token Plan subscription. The preview is available from November 11 to November 25, 2026, giving users a limited period to test the new AI model.
Developers can access Qwen4-Max through the Qwen app, Qoder, QwenWork, and QwenStudio. However, API access is restricted to Token Plan subscribers during the preview. Alibaba has not enabled pay-as-you-go API access for this release.
The model is available through the API under the identifier qwen4-max-preview. It supports a maximum output of 393,216 tokens, covering both reasoning tokens and the final response.
Alibaba has confirmed that the preview model will remain unchanged during the testing period. The company plans to discontinue access to this version after November 25, 2026.
Qwen4-Max Brings New Capabilities for Coding and AI Agent Workflows
Alibaba is positioning Qwen4-Max as an AI model for coding, software development, and multi-step tasks. The model is aimed at professional users who need advanced reasoning and the ability to handle complex instructions. According to Alibaba, Qwen4-Max can support several tasks, including:
- Software development: Writing code and handling complex programming tasks.
- Frontend development: Creating websites and designing user interfaces.
- Office productivity: Processing documents and completing workplace tasks.
- AI agent workflows: Using external tools to perform tasks across different applications.
- Research and data analysis: Reviewing large volumes of information and handling long-context tasks.
These features could make Qwen4-Max useful for developers building AI agents and software applications. Its 1 million-token context window could also help users work with large codebases, lengthy documents, and complex datasets.
However, Alibaba has not yet published independent benchmark results for Qwen4-Max. Its claims about coding, reasoning, and AI agent performance have not been independently verified.
ALSO READ: China May Keep Its Best AI Models at Home and Sell Access Instead
Qwen4-Max Preview Signals the Next Phase of Qwen AI Models
The launch of Qwen4-Max Preview marks an important step in Alibaba’s next generation of Qwen AI models. Its 1 million-token context window and support for image and video inputs put it in competition with other advanced AI models focused on coding, research, and AI agent development.
The preview release also gives developers a chance to test Alibaba’s latest AI model and explore its capabilities before a wider launch.
However, Alibaba has not announced a general release date or final API pricing for Qwen4-Max. The company has also not confirmed whether it plans to release the model’s open weights. A complete technical report and independent benchmark results are also unavailable.
For now, developers will need to test Qwen4-Max to assess its coding, reasoning, and AI agent capabilities. These results will help determine how it compares with competing AI models from other companies.
