Qwen Office launches 10 models to offer personalized AI task solutions
The move shifts office AI from a single-model approach to multi-model orchestration, allowing users to balance performance, speed, and cost.
The short version
- Qwen Office now offers 10 models, including the new Qwen4-Max and third-party models like GLM-5.3 and DeepSeek-V4-Pro.
- The update allows users to choose models based on specific needs like response speed, reasoning depth, and token costs.
- Enterprise administrators can now manage model permissions by department or user group to control AI assets and costs.
- The new Qwen4-Max model features a continuous delivery mechanism that builds content version by version based on user feedback.
Why it matters
This represents a shift toward 'multi-model orchestration' where companies no longer rely on one AI for every task. It allows for more precise cost management and better performance by matching specific tools to specific job requirements.
Moving beyond one-size-fits-all
Qwen Office recently announced a significant expansion of its AI capabilities by launching 10 different large models simultaneously. This move is designed to provide more personalized solutions for both individual users and large enterprises. Previously, the platform was limited to only three tiers of model supply, which often struggled to meet the diverse needs of complex office tasks.
The new lineup includes the recently released Qwen4-Max alongside several prominent third-party models. These include Kimi-K3.1-Pro, GLM-5.3, DeepSeek-V4-Pro, MiniMax-M3.1-Flash, and Step-5-Preview. By opening this list to users, Qwen Office aims to let people choose the specific AI that best fits their unique usage scenarios.
The company explained that different models vary significantly in terms of performance, response speed, and cost. By offering a wider range of choices, the platform intends to help users complete their work at a lower token cost. This shift addresses the growing demand for specialized models as users deploy AI agents into deeper and more specific workflows.
How the new models function
The flagship addition to this lineup is Qwen4-Max, which is described as a new-generation foundation large model. It utilizes a continuous delivery mechanism that allows it to generate a usable initial version of a task first. It then expands that content version by version based on the feedback it receives from the user until the job is finished.
Beyond simple text generation, Qwen4-Max has the specific ability to build operations research models. These are used to solve and verify complex objectives that involve various constraints. This makes the model capable of handling sophisticated technical requirements that standard models might struggle to navigate accurately.
The model can also perform autonomous planning to orchestrate different agents for complex projects. Examples of these types of tasks include website design and scheduling systems for cloud resources. This level of orchestration allows the AI to manage multi-step processes rather than just answering single prompts. Qwen Office provides matching recommendations to help users decide which model is best for their specific task type.
Enterprise governance and control
For corporate clients, the launch includes new model management permissions for administrators. These tools allow leaders to configure which models are available to specific departments, user groups, or even individual members. This ensures that different positions within a company can be matched with the appropriate AI tools based on their work complexity.
This design changes how companies manage their AI assets by moving the choice of a model from an individual act to a matter of organization-level governance. By centralizing these permissions, enterprises can more easily manage their AI resources uniformly across the entire company. It provides a structured way to deploy technology while maintaining oversight.
One of the primary goals of this management system is to avoid cost overruns and compliance risks. In environments with multi-department collaboration, mixed model use can lead to unpredictable spending. By setting clear permissions, administrators can ensure that teams stay within budget and follow internal rules while still having access to the tools they need.
The shift to orchestration
From a broader industry perspective, office AI products are transitioning from a 'one model doing everything' approach to multi-model orchestration. It is difficult for a single vendor's model to be simultaneously superior in reasoning depth, response speed, and price. Aggregating third-party models is becoming a practical way for platforms to fill in capability gaps that a single model might have.
Calling different models for different tasks also helps lower per-token spending, which is a major factor for price-sensitive enterprise procurement. By using a smaller, faster model for simple tasks and a larger, more capable model for complex reasoning, companies can save significant amounts of money. This flexibility is becoming a standard requirement for large-scale AI adoption.
For the end user, having more choices means a greater need for clear judgment on which tool to use. Qwen Office's task classification recommendations serve as a guide for users to navigate these options. For administrators, the success of this multi-model strategy will depend on how smoothly these permissions connect with existing procurement and audit processes.
What to watch
- When Qwen Office will disclose specific billing standards for the 10 models.
- How many enterprises adopt the new department-level permission settings.
- The performance of Qwen4-Max in complex cloud resource scheduling tasks.
Try it
Look for the task classification recommendations in Qwen Office to see which models they suggest for common office workflows.
Sources and links
First reported by BigGo Finance. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.