Microsoft Azure OpenAI

Updated in 2026.19

Use of the Model field for deployment selection in Responses API requests has been deprecated. The removal is planned for Cognigy 2027.3.0. Use the Deployment Name field instead by activating the Use deployment name as model parameter.

As of Cognigy 2026.6.0, the support for the Completions API for custom LLMs is deprecated. The full removal is planned for Cognigy 2026.24.0.

The gpt-5-1, gpt-5-2, gpt-5-4, gpt-5-4-mini, gpt-5-4-nano, gpt-5-5, gpt-5-6-sol, gpt-5-6-terra, and gpt-5-6-luna models default to reasoning.effort: "low" on the Responses API. The reasoning.effort parameter controls how much effort the model uses to think through a request. Higher effort can help the model handle more complex requests, while lower effort can yield faster responses. Supported values, such as "none", "low", "medium", and "high", and their behavior may vary by model and API. Check your LLM provider’s documentation for supported values and behavior.To override this default value, set reasoning.effort in Custom Model Options.The default temperature is 1.0. To override this value, you must set reasoning.effort: "none" and adjust the Temperature parameter or set temperature in Custom Model Options.

To start using a Microsoft Azure OpenAI model with Cognigy features, follow these steps:

Add a Model
Apply the Model

Restrictions

The gpt-5-5, gpt-5-6-sol, gpt-5-6-terra, and gpt-5-6-luna models support only the Responses API.

Add Models

You can add a model using one of the following interfaces:

Add Models via GUI

You can add a model provided by Azure OpenAI to Cognigy in Build > LLM. To add the model, you will need the following parameters:

Standard Model
Custom Model

Parameter	Option	Description
Connection Type		Select one of the following authorization methods: API Key or OAuth2.
	API Key	Enter the Azure API Key. This value can be found in the Keys & Endpoint section when examining your resource from the Azure portal. You can use either `KEY1` or `KEY2`.
	OAuth2	Enter the OAuth2 key. You can use any OAuth2-compliant provider. OAuth 2.0 offers more control and security than API keys by allowing specific permissions, expiring tokens, and reducing exposure through short-lived tokens instead of constant client secret use. To use this type of connection, fill in the following fields: Client ID — enter the Application (client) ID assigned to your app. This ID can be found in the Microsoft Entra admin center: go to App Registrations, select an app, then click Overview. Client Secret — enter the application secret created in the Certificates & secrets section in the Microsoft Entra admin center: go to App Registrations, select an app, then click Overview. OAuth URL — enter the URL to retrieve the access token. This URL points to the `/token` endpoint on the authorization server. The URL can have any format that returns an access token. Scope — add a list of scopes for user permissions, for example, `urn:grp:chatgpt`. Additional Header Name — enter the key for an additional authentication header. This key can have any format, for example, of the OCP APIM Subscription Key. Additional Header Value — enter the value for the additional authentication header. Cognigy uses the following steps to make requests with OAuth 2.0 to an Azure OpenAI model: Requesting a token. Cognigy sends an HTTP POST request to `OAuth URL` with the following parameters: `Client ID`, `Client Secret`, and `Scope`. Receiving the token. The authorization server validates the request and returns an access token to Cognigy. Using the token. When making requests to the Azure OpenAI model, Cognigy includes the received token in the request instead of using an API key.
Resource Name		Enter the resource name. To find this value, go to the Microsoft Azure home page. Under Azure services, click Azure OpenAI. In the left-side menu, under the Azure AI Services section, select Azure OpenAI. Copy the desired resource name from the Name column.
Deployment Name		Enter the deployment name. To find this value, go to the Microsoft Azure home page. Under Azure services, click Azure OpenAI. In the left-side menu, under Azure AI Services, select Azure OpenAI. Select a resource from the Name column. On the resource page, go to Resource Management > Model deployments. On the Model deployments page, click Manage Deployments. On the Deployments page, copy the desired deployment name from the Deployment name column.
API Version		Enter the API version. The API version to use for this operation in the `YYYY-MM-DD` format. Note that the version may have an extended format, for example, `YYYY-MM-DD-preview`.
Custom URL		This parameter is optional. You can use it to route connections between your clusters and the Azure OpenAI provider through dedicated proxy servers for enhanced security. The custom URL can be any valid endpoint URL and must be complete, including any required query parameters. When you specify a custom URL, the Resource Name, Deployment Name, and API Version fields are ignored. For versions 2025.20 and earlier, when selecting the API key connection type, you must use the following pattern: `https://<resource-name>.openai.azure.com/openai/deployments/<deployment-name>/<model-type>?api-version=<api-version>`.
API Type		Select the API format for chat models. Select one of the following API options for chat models: Chat Completion — uses the Chat Completions API. This option is selected by default. Responses — uses the Responses API. For more information, refer to the Microsoft Azure OpenAI documentation.
Use deployment name as model		Appears only when Responses is selected as the API type. When this parameter is activated, Cognigy sends the value of Deployment Name as the model selector in Responses API requests. Microsoft Azure OpenAI deployments require this behavior. When this parameter is deactivated, Cognigy uses the legacy behavior and sends the value of Model instead of Deployment Name. New connections have this parameter active by default. Existing connections have it inactive to preserve backward compatibility.

Parameter	Option	Description
Model Type		Select one of the following options: Chat — lets you use either the Chat Completions API or Responses API for chat models. Completion — uses the Completions API. Embedding — uses the Embedding API. For more information, refer to the Azure OpenAI documentation.
Model Name		Enter the ID of the model that you want to use as a custom model. To find model IDs, refer to the Azure OpenAI documentation.
Connection Type		Select one of the following authorization methods: API Key or OAuth2.
	API Key	Enter the Azure API Key. This value can be found in the Keys & Endpoint section when examining your resource from the Azure portal. You can use either `KEY1` or `KEY2`.
	OAuth2	Enter the OAuth2 key. You can use any OAuth2-compliant provider. OAuth 2.0 offers more control and security than API keys by allowing specific permissions, expiring tokens, and reducing exposure through short-lived tokens instead of constant client secret use. To use this type of connection, fill in the following fields: Client ID — add the Application (client) ID assigned to your app. This ID can be found in the Microsoft Entra admin center: go to App Registrations, select an app, then click Overview. Client Secret — add the application secret created in the Certificates & secrets section in the Microsoft Entra admin center: go to App Registrations, select an app, then click Overview. OAuth URL — add the URL to retrieve the access token. This URL points to the `/token` endpoint on the authorization server. The URL can have any format that returns an access token. Scope — add a list of scopes for user permissions, for example, `urn:grp:chatgpt`. Additional Header Name — enter the key for an additional authentication header. This key can have any format, for example, of the OCP APIM Subscription Key. Additional Header Value — enter the value for the additional authentication header. Cognigy uses the following steps to make requests with OAuth 2.0 to an Azure OpenAI model: Requesting a token. Cognigy sends an HTTP POST request to `OAuth URL` with the following parameters: `Client ID`, `Client Secret`, and `Scope`. Receiving the token. The authorization server validates the request and returns an access token to Cognigy. Using the token. When making requests to the Azure OpenAI model, Cognigy includes the received token in the request instead of using an API key.
Resource Name		Enter the resource name. To find this value, go to the Microsoft Azure home page. Under Azure services, click Azure OpenAI. In the left-side menu, under the Azure AI Services section, select Azure Open AI. Copy the desired resource name from the Name column.
Deployment Name		Enter the deployment name. To find this value, go to the Microsoft Azure home page. Under Azure services, click Azure OpenAI. In the left-side menu, under Azure AI Services, select Azure OpenAI. Select a resource from the Name column. On the resource page, go to Resource Management > Model deployments. On the Model deployments page, click Manage Deployments. On the Deployments page, copy the desired deployment name from the Deployment name column.
API Version		Enter the API version. The API version to use for this operation in the `YYYY-MM-DD` format. Note that the version may have an extended format, for example, `YYYY-MM-DD-preview`.
Custom URL		This parameter is optional. You can use it to route connections between your clusters and the Azure OpenAI provider through dedicated proxy servers for enhanced security. The custom URL can be any valid endpoint URL and must be complete, including any required query parameters. When you specify a custom URL, the Resource Name, Deployment Name, and API Version fields are ignored. For versions 2025.20 and earlier, when selecting the API key connection type, you must use the following pattern: `https://<resource-name>.openai.azure.com/openai/deployments/<deployment-name>/<model-type>?api-version=<api-version>`.
API Type		Appears only when Chat is selected as model type. Select one of the following API options for chat models: Chat Completion — uses the Chat Completions API. This option is selected by default. Responses — uses the Responses API.
Use deployment name as model		Appears only when Responses is selected as the API type. When this parameter is activated, Cognigy sends the value of Deployment Name as the model selector in Responses API requests. Microsoft Azure OpenAI deployments require this behavior. When this parameter is deactivated, Cognigy uses the legacy behavior and sends the value of Model instead of Deployment Name. New connections have this parameter active by default. Existing connections have it inactive to preserve backward compatibility.

Apply changes. Check if the connection was set up by clicking Test. You can switch generative models at any time using the Model list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for Knowledge AI. To switch an embedding, delete the connection and create a new one.

Add Models via the API

You can add either a standard or custom model using the Cognigy API POST /v2.0/largelanguagemodels request. Then, test your connection for the created model via the Cognigy API POST /v2.0/largelanguagemodels//test.

Deprecation of Old Connections for Microsoft Azure OpenAI

As of Cognigy 4.82.0, Azure OpenAI connections that use the AzureOpenAIProvider connection type are deprecated. Cognigy supports the AzureOpenAIProviderV2 connection type instead. If you have old connections (for example, created in the 4.53 release) to Azure OpenAI, these connections have an AzureOpenAIProvider type and are marked with a Deprecated label. Although these connections can still be active, we strongly recommend creating a model with the new AzureOpenAIProviderV2 type, as old connection types will no longer be available in the future. Note that for some Microsoft Azure OpenAI models, such as text-embedding-ada-002 for knowledge search features, you might encounter the following error when an LLM is triggered:

Error while performing knowledge search. Remote returned error: Search failed: Could not fetch embeddings due to missing API resource name for Azure OpenAI

. To resolve the issue, recreate the model and the connection so that both are updated to the latest format.

Apply the Model

To apply a model, follow these steps:

In the left-side menu of the Project, go to Manage > Settings.
Go to the section based on your use case for using a model:
- Generative AI Settings. In the Generative AI Settings section, activate Enable Generative AI Features. This setting is toggled on by default if you have previously set up the Generative AI credentials.
- Knowledge AI Settings. Use this section if you need to add a model for Knowledge AI. Select a model for the Knowledge Search and Answer Extraction features. Refer to the list of standard models and find the models that support these features.
Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select None. Save changes.

Overview

AI Agents

Human Escalation

Platform Features

For Developers

Administration and Installation

Restrictions

Add Models

Add Models via GUI

Add Models via the API

Deprecation of Old Connections for Microsoft Azure OpenAI

Apply the Model

More Information

​Restrictions

​Add Models

​Add Models via GUI

​Add Models via the API

​Deprecation of Old Connections for Microsoft Azure OpenAI

​Apply the Model

​More Information

Restrictions

Add Models

Add Models via GUI

Add Models via the API

Deprecation of Old Connections for Microsoft Azure OpenAI

Apply the Model

More Information