[Product Change] Change announcement for Model Studio US (Virginia) Workspace Deployment Mode
Sep 16, 2026
Alibaba Cloud Model StudioAffected time
2026-10-15, Change
Change and impact
To improve the clarity and manageability of model deployment in the US region and reduce the overhead of operating across multiple regions, Alibaba Cloud Model Studio is upgrading the deployment mode for the US region, effective from the date of this notice. Details are below.
1. Deployment mode upgrade
When you create a workspace, you can choose between two deployment modes — "Global" and "US (Virginia)" — and the inference region is determined by the workspace's deployment mode.
Before | After |
US in-region inference via models with -us suffix | Create a workspace in "US (Virginia)" deployment mode; models are locked to US-only inference |
Global and US models mixed in one workspace, increasing management overhead | Global and US models isolated by workspace, providing clearer separation |
2. Changes to model visibility and usage
Effective from the date of this notice, US in-region inference models will only be available under "US (Virginia)" deployment mode workspaces, without the -us suffix. In "Global" deployment mode workspaces, the original -us suffix models will no longer be displayed, and only non-in-region inference models will be provided. Model names, API protocol, and billing remain unchanged.
- Transition period: If you are currently using -us suffix models, your existing workspace API calls will continue to work during a transition period of approximately one month (until October 15, 2026). Please complete your workspace migration within this period by following the steps in "4. Migration Guide."
- Newly created workspaces do not support calling -us suffix models. Please use model names without the suffix.
3. Models involved
-us | After upgrade |
qwen3.7-max-us | qwen3.7-max |
qwen3.7-plus-us | qwen3.7-plus |
qwen3.6-flash-us | qwen3.6-flash |
wan2.6-i2v-us | wan2.6-i2v |
wan2.6-t2v-us | wan2.6-t2v |
glm-5.2-us | glm-5.2 |
deepseek-v4-pro-0813-us | deepseek-v4-pro-0813 |
deepseek-v4-pro-us | deepseek-v4-pro |
deepseek-v4-flash-0731-us | deepseek-v4-flash-0731 |
deepseek-v4-flash-us | deepseek-v4-flash |
qwen3-vl-flash-us | qwen3-vl-flash |
qwen3-vl-flash-2026-01-22-us | qwen3-vl-flash-2026-01-22 |
qwen3-vl-flash-2025-10-15-us | qwen3-vl-flash-2025-10-15 |
qwen3-asr-flash-us | qwen3-asr-flash |
qwen3-asr-flash-2025-09-08-us | qwen3-asr-flash-2025-09-08 |
qwen-plus-us | qwen-plus |
qwen-plus-2025-12-01-us | qwen-plus-2025-12-01 |
qwen-flash-us | qwen-flash |
qwen-flash-2025-07-28-us | qwen-flash-2025-07-28 |
qwen-mt-lite-us | qwen-mt-lite |
Action required
4. Migration guide
If you are currently using any -us suffix model, follow these steps to migrate:
Step 1 Log in to the Alibaba Cloud Model Studio console and switch to the "US (Virginia)" region.
Step 2 Create a workspace in "US (Virginia)" deployment mode.
Step 3 In that workspace, use the same model name without the `-us` suffix. Capabilities, API protocol, and billing are identical, so no changes to your business logic are required. For example, if you were using qwen3.7-max-us, use qwen3.7-max instead.
Step 4 Repoint the relevant calls from your existing global workspace to the API Key of the new workspace.
If you have any questions, please feel free to contact us via a support ticket.
If you have any questions, please feel free to contact us via our support hotline or by submitting a ticket.