QwenCloud regions, service deployment scopes, Base URLs, and feature availability by region
QwenCloud serves models from multiple regions. Each region has its own endpoint, API keys, workspaces, and model list. Pick a region before you create API keys, and use the endpoint and API key from the same region in your code.
How they relate: you choose a region in the upper-right corner of the console, create a workspace in that region (selecting its service deployment scope where the region offers a choice), and create an API key in that workspace. Your code sends requests to the region's Base URL with that API key. The API key identifies the workspace, so the request runs within that workspace's service deployment scope, and the result is returned through the region.
Examples in this documentation use Singapore (
The service deployment scope determines where inference runs. Germany (Frankfurt), China (Hong Kong), and Japan (Tokyo) use workspaces to separate scopes: choose Global or the local scope when you create a workspace, and API keys from that workspace call models in that scope. Singapore and China North 2 (Beijing) need no selection.
In Singapore, inference runs on nodes outside the Chinese mainland.
If you have data compliance requirements, choose a scope with a specific geographic boundary. Otherwise, choose Global for a larger pool of inference resources. Model availability and prices can differ between scopes in the same region. See Model prices.
The default workspace in Frankfurt, China (Hong Kong), and Japan (Tokyo) uses the Global scope. A workspace's scope is set when the workspace is created and cannot be changed afterward.
The service deployment scope does not change the Base URL.
All regions use the domain format
Append the path for the protocol you use. Singapore examples:
Use a pay-as-you-go API key created in the same region. A key from one region cannot call the endpoint of another region. See API keys.
Models differ by region. Before you call a model in a region other than Singapore, check that the model is listed for that region in Model prices.
Token Plan is available only in Singapore and uses its own endpoint and API key. It does not use the region domains above.
See Token Plan.
Menus for features that are not supported in the current region are hidden in the console. To use them, switch to a region that supports them.
Use the region selector in the upper-right corner of the console to view or change the current region.
The console opens in Singapore by default. The region you select applies to all console pages and stays selected when you move between pages. Switch it again in the region selector to see another region's resources.
When you switch regions:
Try AI uses the models available in the current region. If a scene category (text, image, or video) has no available model in the current region, its entry is hidden.
When you switch regions in Try AI:
The API key and the Base URL are from different regions, or a Token Plan key is used with a pay-as-you-go endpoint (or the reverse). Use an API key created in the same region as the Base URL. See Error codes.
The model is not offered in the region you are calling. Check the region's model list in Model prices, or call the model from a region that offers it.
They belong to the region where you created them. Switch back to that region in the console.
The feature is not supported in the current region. See Feature availability by region and switch to a supported region.
Billing pages are not filtered by region or workspace. Regional usage is on Analytics > Usage, which follows the selected region. See Bill query.
Key concepts
| Concept | What it determines |
|---|---|
| Region | Where your requests are received and where your data is stored. Each region has its own endpoint, API keys, workspaces, and model list. |
| Workspace | The unit for organizing API keys, model access, and rate limits within a region. Each region has its own workspaces. See Workspaces. |
| Service deployment scope | Where model inference runs. In Germany (Frankfurt), China (Hong Kong), and Japan (Tokyo), you choose it per workspace. |
| Base URL | The endpoint your code sends requests to. It contains only the region ID, for example ap-southeast-1.maas.qwencloudapi.com. It does not contain a workspace ID. |
Supported regions
| Region | Region ID |
|---|---|
| Singapore | ap-southeast-1 |
| China North 2 (Beijing) | cn-beijing |
| Germany (Frankfurt) | eu-central-1 |
| China (Hong Kong) | cn-hongkong |
| Japan (Tokyo) | ap-northeast-1 |
ap-southeast-1). To use another region, replace ap-southeast-1 in the Base URL and use an API key created in that region.
Choose a region close to your users to reduce latency. If you have data residency requirements, choose the region where your data must be stored.
Service deployment scopes
The service deployment scope determines where inference runs. Germany (Frankfurt), China (Hong Kong), and Japan (Tokyo) use workspaces to separate scopes: choose Global or the local scope when you create a workspace, and API keys from that workspace call models in that scope. Singapore and China North 2 (Beijing) need no selection.
| Region | Service deployment scope | Inference location |
|---|---|---|
| Germany (Frankfurt) | Global | Any available node, including the Chinese mainland and other regions |
| EU | The European Union only | |
| China (Hong Kong) | Global | Any available node, including the Chinese mainland and other regions |
| China (Hong Kong) | China (Hong Kong) only | |
| Japan (Tokyo) | Global | Any available node, including the Chinese mainland and other regions |
| Japan | Japan only |
Endpoints
Pay-as-you-go
All regions use the domain format {region}.maas.qwencloudapi.com, where {region} is the region ID.
| Region | Domain |
|---|---|
| Singapore | ap-southeast-1.maas.qwencloudapi.com |
| China North 2 (Beijing) | cn-beijing.maas.qwencloudapi.com |
| Germany (Frankfurt) | eu-central-1.maas.qwencloudapi.com |
| China (Hong Kong) | cn-hongkong.maas.qwencloudapi.com |
| Japan (Tokyo) | ap-northeast-1.maas.qwencloudapi.com |
| Protocol | Base URL (Singapore) |
|---|---|
| OpenAI compatible (Chat Completions, Responses) | https://ap-southeast-1.maas.qwencloudapi.com/compatible-mode/v1 |
| Anthropic compatible | https://ap-southeast-1.maas.qwencloudapi.com/apps/anthropic |
| DashScope HTTP | https://ap-southeast-1.maas.qwencloudapi.com/api/v1 |
| DashScope WebSocket | wss://ap-southeast-1.maas.qwencloudapi.com/api-ws/v1/inference |
| Realtime (WebSocket) | wss://ap-southeast-1.maas.qwencloudapi.com/api-ws/v1/realtime |
maas.qwencloudapi.com (without a region ID) is the Singapore endpoint and is equivalent to ap-southeast-1.maas.qwencloudapi.com. Existing code that uses it continues to call Singapore and does not need to change.Token Plan
Token Plan is available only in Singapore and uses its own endpoint and API key. It does not use the region domains above.
| Protocol | Base URL |
|---|---|
| OpenAI compatible | https://token-plan.maas.qwencloudapi.com/compatible-mode/v1 |
| Anthropic compatible | https://token-plan.maas.qwencloudapi.com/apps/anthropic |
Feature availability by region
| Feature | Singapore | China North 2 (Beijing) | Germany (Frankfurt) | China (Hong Kong) | Japan (Tokyo) |
|---|---|---|---|---|---|
| Model calls (real-time) | Supported | Supported | Supported | Supported | Supported |
| Batch | Supported | Supported | Not supported | Not supported | Not supported |
| Try AI | Supported | Supported | Supported | Supported | Supported |
| API keys | Supported | Supported | Supported | Supported | Supported |
| Workspaces | Supported | Supported | Supported | Supported | Supported |
| Datasets | Supported | Supported | Not supported | Not supported | Not supported |
| Fine-tuning | Supported | Supported | Not supported | Not supported | Not supported |
| Evaluation | Supported | Supported | Not supported | Not supported | Not supported |
| Deployment | Supported | Supported | Not supported | Not supported | Not supported |
| Alerts | Supported | Not supported | Not supported | Not supported | Not supported |
| Token Plan | Supported | Not supported | Not supported | Not supported | Not supported |
| Free quota | Supported | Not supported | Not supported | Not supported | Not supported |
| Billing and usage | Supported | Supported | Supported | Supported | Supported |
Switch regions
Use the region selector in the upper-right corner of the console to view or change the current region.
The console opens in Singapore by default. The region you select applies to all console pages and stays selected when you move between pages. Switch it again in the region selector to see another region's resources.
When you switch regions:
- The region applies to all pages. Every console page uses the selected region until you switch again.
- Resources stay in their original region. Workspaces, API keys, and other resources are not copied to the new region and are not deleted. To view them, switch back to the original region.
- Model list, prices, and rate limits change to those of the new region.
- Your code does not change. Switching regions in the console does not change the Base URL or API key in your local code. To call another region, update both in your code.
- Billing pages do not change. Billing Overview, Order Details, and Pay-as-you-go cover all regions and are not filtered by region or workspace. This includes available balance and recent spend on the console home page. The token usage trend follows the selected region, like Analytics > Usage.
Try AI and regions
Try AI uses the models available in the current region. If a scene category (text, image, or video) has no available model in the current region, its entry is hidden.
When you switch regions in Try AI:
- Unsent text is not kept, and attachments and parameters are cleared.
- Completed conversations remain available in History.
- A conversation that has not completed, including one that is still generating, is lost and cannot be viewed.
- The selected model stays selected if it is available in the new region.