Qwen3.8-Max-Preview - The latest generation base model launched by Alitongyi

What is Qwen3.8-Max-Preview?

Qwen3.8-Max-Preview is a preview version of the latest generation of large-scale foundational model launched by Alibaba's Tongyi Qianwen platform. With 2.4TB of parameters, it has been deployed on Qoder, Token Plan, and QoderWork platforms. The model has demonstrated capabilities comparable to world-leading models in complex multi-agent long-term tasks such as full-stack development, data analysis, and Office workflows. The official version will be released and open-sourced soon.

Main functions of Qwen3.8-Max-Preview

• Long context processingIt supports three context window configurations: 200K, 400K, and 1M, to meet the needs of long document comprehension and long-term dialogue.
• Deep thinking modeBuilt-in thinking mode switch, which can automatically switch between deep reasoning and fast response according to the complexity of the task.
• Multi-agent collaborationIt supports complex task decomposition and multi-agent collaboration, covering scenarios such as full-stack development, data analysis, and office automation.
• Full-modal capabilityUnder the unified subscription system, it can access full-modal AI capabilities such as text, voice, and images.
• Compatible with mainstream toolsSupports OpenAI / Anthropic protocols and can be seamlessly integrated with mainstream coding/agent tools such as Cursor and Windsurf.

The technical principles of Qwen3.8-Max-Preview

• Ultra-large-scale MoE architectureIt adopts a hybrid expert architecture with a total of 2.4T parameters, and releases powerful performance while maintaining inference efficiency through a sparse activation mechanism.
• Continuous Evolution TrainingThe model is continuously iterated and optimized on a daily basis, achieving significant improvements over Qwen3.7-Max in internal real-world task evaluations.
• Long-range dependency modelingBy expanding the context window to the million level, it enhances the semantic understanding of ultra-long texts and the ability to associate information across paragraphs.

How to use Qwen3.8-Max-Preview

The Qwen 3.8-Max preview model has been released online.Alibaba Token PlanQoder and QoderWork.

The core advantages of Qwen3.8-Max-Preview

• Top performanceThe official claim is that the model may be the most powerful model besides Fable 5, and it can rival the world's cutting-edge level in complex real-world tasks.
• Open source commitmentAfter the preview version is available for early access, the official version will be released and open source, taking into account both commercial applications and the community ecosystem.
• Flexible subscriptionToken Plan offers multiple packages for individuals and teams, supporting 1 to 8 agents running concurrently.

Comparison of Qwen3.8-Max-Preview with similar competing products

Application scenarios of Qwen3.8-Max-Preview

• Complex software developmentUse multiple agents to collaborate on full-stack project planning, code writing, debugging, and documentation generation.
• Massive data analysis: Use a 1M context window to process extremely long reports and log files, extract key insights, and generate visual conclusions.
• Smart office automationIt integrates with Office workflows to automatically complete tasks such as email composition, PPT creation, and meeting minutes organization.
• In-depth reading of long documentsIt can perform abstracting, Q&A, and cross-chapter correlation analysis on extremely long texts such as academic papers, legal contracts, and technical manuals.
• AI-assisted creationIt serves as the underlying model in coding/agent tools, providing developers with real-time code completion and architecture suggestions.