Alibaba's QwenWork Adds Kimi, GLM, DeepSeek, MiniMax and Step Models, With Admin-Level Access Controls
Alibaba's QwenWork office agent opened a broader model menu on November 16, adding third-party models from Kimi, GLM, DeepSeek, MiniMax and Step next to its own Qwen models. Enterprise administrators can set model access by department, user group or individual member.
Alibaba's QwenWork office agent now lets users choose third-party models from Kimi, GLM, DeepSeek, MiniMax and Step alongside Alibaba's own Qwen models. The expanded menu opened on November 16 and is available to personal and enterprise standard-edition users, according to Alibaba's announcement as reported by Leiphone, a Chinese-language outlet. Pandaily also covered the announcement, and the company's own QwenWork site is in Chinese.
What is in the menu
Users can now pick the model behind an office task instead of relying only on Alibaba's existing standard, advanced and Qwen3.8-Max options. The published menu includes:
- Kimi-K3.1-Pro
- GLM-5.3
- GLM-5.3-Flash
- DeepSeek-V4-Pro
- DeepSeek-V4.1-Flash
- MiniMax-M3.1-Flash
- Step-5-Preview
- Qwen4-Max, which Alibaba describes as newly released
These are options inside QwenWork. The update is not a joint model release by the companies represented in the menu.
Alibaba says the expansion responds to differences in reasoning ability, response speed and token consumption as people hand agents more specialized work.
Alibaba's task guidance
Alibaba recommends Qwen4-Max for work that involves multiple steps of complex reasoning. The company says the model can deliver an initial usable version of an output, then expand and revise it in response to feedback. It also describes the model as planning, solving and checking work against complex goals and constraints, citing website design and cloud-resource scheduling as examples. These are Alibaba's capability claims, not results of an independent comparison.
For organizing rules, reviewing plans, running simulations and building 3D presentations, Alibaba points users toward GLM-5.3 and DeepSeek-V4-Pro.
For faster everyday work such as drafting copy, extracting information and organizing content, the company recommends lower-priced options: Qwen3.8-Flash, GLM-5.3-Flash and DeepSeek-V4.1-Flash.
Alibaba's announcement does not provide a common benchmark establishing which option is best for every task. Per-model pricing, context window sizes and benchmark scores for the menu options are not yet disclosed.
Enterprise controls
Enterprise administrators can configure model access by department, user group or individual member, according to Alibaba. An organization can therefore offer different model choices to different roles instead of exposing one menu to everyone.
Context
The update follows QwenWork's recent 2.0 team-workspace release. The immediate change sits at the model-selection and permissions layer: the office agent can draw on a wider set of providers, while the user or organization decides which models are available for a given job.
What this means
QwenWork now treats the model as a per-task choice rather than a fixed backend. Alibaba's own guidance frames the decision around three variables: reasoning depth, response speed and token consumption. Its recommendations split along those lines, with Qwen4-Max for multi-step reasoning, GLM-5.3 and DeepSeek-V4-Pro for structured analysis and presentation work, and the Flash-tier models for routine drafting and extraction.
The department, group and individual-level access settings give administrators a way to limit which models staff can use. Those recommendations are Alibaba's, and with no shared benchmark published, users have no common yardstick for comparing the options on their own workloads.
Related Articles
Doubao Work launches workbenches with preset skills and tools for 21 professional scenarios
ByteDance's Doubao Work launched a workbench feature on November 16 that bundles preconfigured skills and tools for 21 professional scenarios. It is available in the Doubao Work desktop client and the Doubao desktop app. Users can also build, save and share their own workspaces.
Google Adds Shareable Interactive Sites to Gemini Notebook, Rolling Out Globally on Web
Google announced on November 16, 2026 that Gemini Notebook can turn a notebook's sources into interactive sites. Outputs now include custom visual explainers, dashboards and project trackers, and the feature is rolling out globally on the web.
Perplexity launches Monitor API: web monitoring priced at $1 per 1,000 successful runs
Perplexity launched the Monitor API on November 16, 2026. It gives developers scheduled topic searches and webpage-change tracking, with filtered updates delivered through the API or signed webhooks. Web monitoring costs $1 per 1,000 successful runs.
Siri AI can now answer Health data questions in iOS 27.3 beta 2
Apple's iOS 27.3 beta 2, released to developers on November 16, lets Siri AI answer questions about Health app data. The assistant had been blocked from that data since Siri AI launched in iOS 27.0, while the older Siri Classic could access it.
Comments
Loading...