Agent operation depends on model configuration. The model management feature supports connecting to multiple AI model vendors, centrally managing API endpoints and authentication information, and verifying model availability before formal use.
Add a model connection
-
Log on to the AgentCore console.
-
In the left-side navigation pane of the details page, click Model Connections.
-
Click the Model List tab.
-
Click Add Model Connection.
-
Configure the basic model information.
-
Select a preset or custom model vendor.
-
Enter a model connection name.
-
If you select a custom model vendor, configure the API endpoint.
-
Enter the access credential. Currently, only manual entry of API Key is supported.
-
-
Under Model Configuration, select the required models.
-
Click OK to complete adding the model.
Manage models
View model details
Click Details to view the basic model information.
Basic information
|
Field |
Description |
|
Model name |
The name of the model, for example, qwen3.8-max. |
|
Model ID |
The unique identifier assigned by the system to the model. |
|
Model connection |
The model connection ID that the model belongs to. |
|
Model vendor |
The vendor of the model, for example, Alibaba Cloud Model Studio (Bailian). |
|
Protocol |
The API protocol used for model invocation, for example, OpenAI/v1. |
|
Created at |
The time when the model was connected. |
Model capabilities
Displays the capabilities supported by the model. Each item is marked as supported or not supported.
|
Capability |
Description |
|
Tool use |
Whether the model supports using tools. |
|
Multiple tool use |
Whether the model supports using multiple tools in a single request. |
|
Streaming tool use |
Whether the model supports using tools in streaming mode. |
|
Vision |
Whether the model supports image input. |
|
Audio |
Whether the model supports audio input. |
|
Video |
Whether the model supports video input. |
|
Documents |
Whether the model supports document input. |
|
Reasoning |
Whether the model supports reasoning. |
Invocation parameters
|
Field |
Description |
|
Context window |
The maximum number of tokens that the model can process in a single request. |
|
Max output tokens |
The maximum number of tokens that the model can output in a single response. |
|
Streaming output |
Whether the model supports streaming output. |
Debug model
Click Debug to test the model output.
Delete model
Click Delete to delete the model.
Manage model connections
View model details
Click Details to view the basic model information.
Basic information
|
Field |
Description |
|
Model vendor |
The vendor connected by the model connection, for example, Alibaba Cloud Model Studio (Bailian). |
|
API endpoint |
The model invocation address provided by the vendor, for example, v1 · dashscope.aliyuncs.com. |
|
Protocol |
The API protocol used for model invocation, for example, OpenAI/v1. |
Model configuration
The model configuration area displays the models connected under this connection, grouped by model type. The number after the group label indicates the number of models in that type. You can switch between card view and list view in the upper-right corner of the area.
Each model displays the following information:
|
Field |
Description |
|
Model name |
The name of the model, for example, qwen3.8-max. |
|
Context window |
The maximum number of tokens that the model can process in a single request, for example, 1m indicates 1 million tokens. |
|
Model capabilities |
The capability tags supported by the model, such as Tool use, Multiple tool use, Streaming tool use, Vision, Video, and Reasoning. Unsupported capabilities do not display tags. |
Click the model name to go to the model details page and view the model ID, creation time, and complete capabilities and invocation parameters.
Debug model
Click Debug to test the model connection output.
Delete model
Click Delete to delete the model connection.