Deploy a model by using an API or CLI - Alibaba Cloud Model Studio

Prerequisites

You have read the Model Deployment Overview and are familiar with the supported models and basic deployment steps.
Obtain an API key and export the API key as an environment variable.

1. Deploy the model

The following command uses the fine-tuned custom model qwen3-8b-ft-202511132025-0260 to create a dedicated service named qwen3-8b-ft-202511132025-0260. To obtain the custom model ID, go to the Model Studio console - model finetuning page. Click the Task Name of the model you want to deploy, click Outputs, and then click the blue model name. This opens the My Models page, where you can find the Model ID in the basic information section. Use the Model IDas the input for the model_name parameter to deploy the model by using the API.

Provisioned Throughput (PTU)
Model unit
Token usage

After you run the deployment command, billing starts as soon as the service is successfully deployed, even if you do not use it. Before proceeding, we recommend you review the service billing rules.

The provisioned throughput billing method charges based on usage duration. This method is suitable for scenarios that require stable throughput, high concurrency, low latency, and predictable traffic. In this mode, the platform provisions both throughput/concurrency and generation speed, which you cannot adjust.

curl "https://dashscope-intl.aliyuncs.com/api/v1/deployments" \
--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
--header 'Content-Type: application/json' \
--data '{
    "name": "my_qwen_flash",
    "model_name": "qwen-flash-2025-07-28",
    "plan": "ptu",
    "ptu_capacity": {
        "input_tpm": 10000,
	"output_tpm": 1000
    }
}'

After you run the deployment command, billing starts as soon as the service is successfully deployed, even if you do not use it. Before proceeding, we recommend you review the service billing rules.
Computing resources for the post-paid model unit plan are allocated on a first-come, first-served basis. If the purchase is unsuccessful, a full refund will be issued.

The model unit billing method charges you based on usage duration. This billing method is ideal for large-scale inference tasks after model finetuning, offering dedicated resources with flexible performance and cost adjustments. You can customize both throughput/concurrency and generation speed.

curl "https://dashscope-intl.aliyuncs.com/api/v1/deployments" \
--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
--header 'Content-Type: application/json' \
--data '{
    "name": "my_qwen_plus",
    "model_name": "qwen-plus-2025-12-01",
    "plan": "mu",
    "deploy_spec": "MU1",
    "enable_thinking": true,
    "capacity": 4,
    "max_context_length": 10000,
    "rpm_limit": 500,
    "tpm_limit": 1000
}'

The model unit deployment mode supports the following additional settings:

Configuration item	Configuration details
Service name	Custom name for the deployment service.
Select model	Select the model to deploy, including platform built-in models and fine-tuned models.
Model unit type	Select the deployment specification. Different specifications correspond to different compute power and performance.
Deployment replica count	Set the initial number of deployment replicas, which affects the concurrency capacity of the service.
Deployment template	Select a deployment template (such as "single-node deployment"). Different templates correspond to different resource configuration schemes. Available only under the model unit billing mode.
Configure model inference mode	For some models deployed in Model Unit mode, you can configure the inference mode, maximum context length, and more. Instruct - After deployment, the model performs inference in non-thinking mode. Thinking - After deployment, the model performs inference in thinking mode.
Maximum context length	The Model Unit deployment mode of some models supports this setting. The maximum context length depends on the model type.
Service rate limiting	The Model Unit deployment mode of some models supports this setting. You can limit the RPM and TPM of model invocations.

To learn how to configure these settings by using the API, see Create a model deployment task by using an API.

With token usage billing, you are charged based on token usage. This method is suitable for cost-sensitive scenarios where concurrency and latency requirements are not critical. This mode offers the best price advantage; the platform provisions throughput/concurrency and generation speed, which you cannot adjust.

curl "https://dashscope-intl.aliyuncs.com/api/v1/deployments" \
--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
--header 'Content-Type: application/json' \
--data '{
    "model_name": "qwen3-8b-ft-202511132025-0260",
    "plan": "lora",
    "capacity": 1,
    "name": "qwen3-8b-ft"
}'

The capacity parameter is required but currently has no effect. To request scaling, go to the model deployment console and submit a form.

On success, the command returns the following result (using a LoRA deployment as an example):

{
    "request_id": "83b173ab-2b2f-41aa-8c57-b173e8be934e",
    "output":
    {
        "deployed_model": "qwen3-8b-ft-202511132025-0260",
        "gmt_create": "2025-11-20T20:06:46.405",
        "gmt_modified": "2025-11-20T20:06:46.405",
        "status": "PENDING",
        "model_name": "qwen3-8b-ft-202511132025-0260",
        "base_model": "qwen3-8b",
        "workspace_id": "llm-8v*****",
        "charge_type": "post_paid",
        "creator": "16542*****",
        "modifier": "16542*****",
        "plan": "***"
    }
}

Where deployed_model is the unique ID of the dedicated service.

2. Query service status

Run the following command to query the details of a specific dedicated service:

curl "https://dashscope-intl.aliyuncs.com/api/v1/deployments/qwen3-8b-ft-202511132025-0260" \
    --header "Authorization: Bearer $DASHSCOPE_API_KEY" \
    --header 'Content-Type: application/json'

On success, the command returns the following result:

{
    "request_id": "ca36952d-9136-426e-ab08-68a97ad72719",
    "output":
    {
        "deployed_model": "qwen3-8b-ft-202511132025-0260",
        "gmt_create": "2025-11-20T20:32:08",
        "gmt_modified": "2025-11-20T20:42:25",
        "status": "RUNNING",
        "model_name": "qwen3-8b-ft-202511132025-0260",
        "base_model": "qwen3-8b",
        "base_capacity": 2,
        "capacity": 2,
        "ready_capacity": 2,
        "workspace_id": "llm-8v53etv3hwb8orx1",
        "charge_type": "post_paid",
        "creator": "1654290265984853",
        "modifier": "1654290265984853",
        "plan": "mu",
        "model_unit_spec": "MU1"
    }
}

When the service status is RUNNING, the service deployment is complete.

3. Make inference requests

If this is your first time using the DashScope SDK, see Install the SDK.Ensure the API Key's workspace matches the model's deployment workspace.

Use the SDK to send a request to the dedicated service:

from dashscope import Generation
from http import HTTPStatus
import os
response = Generation.call(
    model='qwen3-8b',
    prompt='Who are you?',
    enable_thinking=False,
    api_key=os.getenv('DASHSCOPE_API_KEY'),
)
if response.status_code == HTTPStatus.OK:
    print(response.output)
    print(response.usage)
else:
    print(response.code)
    print(response.message)

A successful execution returns the following result:

{"text": "I am Qwen, a large language model developed by Alibaba Cloud. I am designed to generate various types of text, such as articles, stories, and poems, and to engage in conversations, answer questions, provide information, and offer help in different scenarios. I'm happy to serve you! If you have any questions or need assistance, please feel free to let me know.", "finish_reason": "stop", "choices": null}
{"input_tokens": 11, "output_tokens": 63, "total_tokens": 74}

4. Delete the dedicated service

Running the delete command immediately takes the deployment service offline. This action is irreversible and has the following effects:

You can no longer call the model.
Billing for the service stops.

If you no longer need a dedicated service, you can delete it by using the following command:

curl --request DELETE 'https://dashscope-intl.aliyuncs.com/api/v1/deployments/qwen3-8b-ft-202511132025-0260' \
    --header "Authorization: Bearer $DASHSCOPE_API_KEY" \
    --header 'Content-Type: application/json'

On success, the command returns the following result:

{
    "request_id": "8f726017-6042-420e-a465-0d366a3aba59",
    "output":
    {
        "deployed_model": "qwen3-8b-ft-202511132025-0260",
        "gmt_create": "2025-11-20T20:32:08",
        "gmt_modified": "2025-11-27T16:35:31.591",
        "status": "DELETING",
        "model_name": "qwen3-8b-ft-202511132025-0260",
        "base_model": "qwen3-8b",
        "base_capacity": 2,
        "capacity": 2,
        "ready_capacity": 2,
        "workspace_id": "llm-8v53etv3hwb8orx1",
        "charge_type": "post_paid",
        "creator": "1654290265984853",
        "modifier": "1654290265984853",
        "plan": "mu",
        "model_unit_spec": "MU1"
    }
}

After deleting the service, you can no longer query its status using the method described in 2. Query service status.

API reference

For details on the API calls, see API details.

FAQ

Permission errors during deployment

When deploying a model by using an API, ensure the following:

The API Key's Workspace must have permission to manage the model. Go to the Business Space Management page in Model Studio and check the model deployment permission settings for the corresponding workspace.
API call error: Workspace xxx does not have deployment privilege for model xxxx.
In the Actions column for the corresponding workspace, click Model Permission and Flow Control Settings. In the Model List, find the target model and check the authorization status in the Model Deployment column. If it shows Not Authorized, click Edit in the Actions column to grant permission. If you are prompted about insufficient permissions, contact your organization or IT administrator to add the relevant permissions or perform the operation on your behalf.
The Owner Account that owns the API Key has the required permissions in the Workspace. Go to the Model Studio console, click the workspace in the lower-left corner, switch to the correct workspace, and then click to check the model deployment permission settings.
API call error: Workspace access denied.
In the left-side navigation pane, click Permission Management and confirm that the user list includes the API Key's owner account (with type primary account). If you are prompted about insufficient permissions, contact your organization or IT administrator to add the relevant permissions or perform the operation on your behalf.

​Prerequisites

​1. Deploy the model

​2. Query service status

​3. Make inference requests

​4. Delete the dedicated service

​API reference

​FAQ

​Permission errors during deployment

Prerequisites

1. Deploy the model

2. Query service status

3. Make inference requests

4. Delete the dedicated service

API reference

FAQ

Permission errors during deployment