Azure OpenAI Service models - Azure OpenAI

GPT-4

GPT-4 can solve difficult problems with greater accuracy than any of OpenAI's previous models. Like GPT-3.5 Turbo, GPT-4 is optimized for chat and works well for traditional completions tasks. Use the Chat Completions API to use GPT-4. To learn more about how to interact with GPT-4 and the Chat Completions API check out our in-depth how-to .

To request access to GPT-4, Azure OpenAI customers can apply by filling out this form


    gpt-4


    gpt-4-32k

The gpt-4 model supports 8192 max input tokens and the gpt-4-32k model supports up to 32,768 tokens.

GPT-3.5

GPT-3.5 models can understand and generate natural language or code. The most capable and cost effective model in the GPT-3.5 family is GPT-3.5 Turbo, which has been optimized for chat and works well for traditional completions tasks as well. We recommend using GPT-3.5 Turbo over legacy GPT-3.5 and GPT-3 models .


    gpt-35-turbo


    gpt-35-turbo-16k

The gpt-35-turbo model supports 4096 max input tokens and the gpt-35-turbo-16k model supports up to 16,384 tokens.

Like GPT-4, use the Chat Completions API to use GPT-3.5 Turbo. To learn more about how to interact with GPT-3.5 Turbo and the Chat Completions API check out our in-depth how-to .

Embeddings models

Important

We strongly recommend using text-embedding-ada-002 (Version 2) . This model/version provides parity with OpenAI's text-embedding-ada-002 . To learn more about the improvements offered by this model, please refer to OpenAI's blog post . Even if you are currently using Version 1 you should migrate to Version 2 to take advantage of the latest weights/updated token limit. Version 1 and Version 2 are not interchangeable, so document embedding and document search must be done using the same version of the model.

The previous embeddings models have been consolidated into the following new replacement model:

text-embedding-ada-002

DALL-E (Preview)

The DALL-E models, currently in preview, generate images from text prompts that the user provides.

Whisper (Preview)

The Whisper models, currently in preview, can be used for speech to text.

You can also use the Whisper model via Azure AI Speech batch transcription API. Check out What is the Whisper model? to learn more about when to use Azure AI Speech vs. Azure OpenAI Service.

Model summary table and region availability

Important

Due to high demand:

South Central US is temporarily unavailable for creating new resources and deployments.

GPT-4 models

These models can only be used with the Chat Completion API.

Model ID Base model Regions Fine-Tuning Regions Max Request (tokens) Training Data (up to)


   gpt-4

¹ ³ (0613) Australia East, Canada East, East US, East US 2, France Central, Japan East, Sweden Central, Switzerland North, UK South 8,192 September 2021


   gpt-4-32k

¹ ³ (0613) Australia East, Canada East, East US, East US 2, France Central, Japan East, Sweden Central, Switzerland North, UK South 32,768 September 2021

¹ The model is only available by request .
² Version 0314 of gpt-4 and gpt-4-32k will be retired no earlier than July 5, 2024. See model updates for model upgrade behavior.
³ We are rolling out availability of new regions to customers gradually to ensure a smooth experience. In East US and France Central, customers with existing deployments of GPT-4 can create additional deployments of GPT-4 version 0613. For customers new to GPT-4 on Azure OpenAI, please use one of the other available regions.

GPT-3.5 models

GPT-3.5 Turbo is used with the Chat Completion API. GPT-3.5 Turbo (0301) can also be used with the Completions API. GPT3.5 Turbo (0613) only supports the Chat Completions API.

Model ID Base model Regions Fine-Tuning Regions Max Request (tokens) Training Data (up to)


   gpt-35-turbo

(0613) Australia East, Canada East, East US, East US 2, France Central, Japan East, North Central US, Sweden Central, Switzerland North, UK South 4,096 Sep 2021


   gpt-35-turbo-16k

(0613) Australia East, Canada East, East US, East US 2, France Central, Japan East, North Central US, Sweden Central, Switzerland North, UK South 16,384 Sep 2021

¹ Version 0301 of gpt-35-turbo will be retired no earlier than July 5, 2024. See model updates for model upgrade behavior.

Embeddings models

These models can only be used with Embedding API requests.

text-embedding-ada-002 (version 2) Canada East, East US, East US2, France Central, Japan East, North Central US, South Central US, Switzerland North, UK South, West Europe 8,191 Sep 2021 text-embedding-ada-002 (version 1) East US, South Central US, West Europe 2,046 Sep 2021

DALL-E models (Preview)

Model ID Base model Regions Fine-Tuning Regions Max Request (characters) Training Data (up to)

Working with models

Finding what models are available

You can get a list of models that are available for both inference and fine-tuning by your Azure OpenAI resource by using the Models List API .

Model updates

Azure OpenAI now supports automatic updates for select model deployments. On models where automatic update support is available, a model version drop-down will be visible in Azure OpenAI Studio under Create new deployment and Edit deployment :

Auto update to default

When Auto-update to default is selected your model deployment will be automatically updated within two weeks of a change in the default version.

If you are still in the early testing phases for inference models, we recommend deploying models with auto-update to default set whenever it is available.

Specific model version

As your use of Azure OpenAI evolves, and you start to build and integrate with applications you may want to manually control model updates so that you can first test and validate that model performance is remaining consistent for your use case prior to upgrade.

When you select a specific model version for a deployment this version will remain selected until you either choose to manually update yourself, or once you reach the retirement date for the model. When the retirement date is reached the model will auto-upgrade to the default version at the time of retirement.

GPT-35-Turbo 0301 and GPT-4 0314 retirement

The gpt-35-turbo ( 0301 ) and both gpt-4 ( 0314 ) models will be retired no earlier than July 5, 2024. Upon retirement, deployments will automatically be upgraded to the default version at the time of retirement. If you would like your deployment to stop accepting completion requests rather than upgrading, then you will be able to set the model upgrade option to expire through the API. We will publish guidelines on this by September 1.

Viewing deprecation dates

For currently deployed models, from Azure OpenAI Studio select Deployments :


   OnceNewDefaultVersionAvailable

Once a new version is designated as the default, the model deployment will auto-upgrade to the default version within two weeks of that designation change being made.


   OnceCurrentVersionExpired

Once the retirement date is reached the model deployment will auto-upgrade to the current default version.


   NoAutoUpgrade

The model deployment will never auto-upgrade. Once the retirement date is reached the model deployment will stop working. You will need to update your code referencing that deployment to point to a non-expired model deployment.

To query the current model deployment settings including the deployment upgrade configuration for a given resource use Deployments List

GET https://management.azure.com/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.CognitiveServices/accounts/{accountName}/deployments?api-version=2023-05-01
Path parameters
Parameter
Required?
Description
Example response
      "id": "/subscriptions/{Subcription-GUID}/resourceGroups/{Resource-Group-Name}/providers/Microsoft.CognitiveServices/accounts/{Resource-Name}/deployments/text-davinci-003",
      "type": "Microsoft.CognitiveServices/accounts/deployments",
      "name": "text-davinci-003",
      "sku": {
        "name": "Standard",
        "capacity": 60
      "properties": {
        "model": {
          "format": "OpenAI",
          "name": "text-davinci-003",
          "version": "1"
        "versionUpgradeOption": "OnceNewDefaultVersionAvailable",
        "capabilities": {
          "completion": "true",
          "search": "true"
        "raiPolicyName": "Microsoft.Default",
        "provisioningState": "Succeeded",
        "rateLimits": [
            "key": "request",
            "renewalPeriod": 10,
            "count": 60
            "key": "token",
            "renewalPeriod": 60,
            "count": 60000
You can then take the settings from this list to construct an update model REST API call as described below if you want to modify the deployment upgrade configuration.
Update & deploy models via the API
PUT https://management.azure.com/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.CognitiveServices/accounts/{accountName}/deployments/{deploymentName}?api-version=2023-05-01
Path parameters
Parameter
Required?
Description
string
Required
The deployment name you chose when you deployed an existing model or the name you would like a new model deployment to have.
resourceGroupName
string
Required
The name of the associated resource group for this model deployment.
subscriptionId
string
Required
Subscription ID for the associated subscription.
api-version
string
Required
The API version to use for this operation. This follows the YYYY-MM-DD format.
Supported versions
2023-05-01 Swagger spec
Request body
This is only a subset of the available request body parameters. For the full list of the parameters, you can refer to the REST API reference documentation.
Parameter
Description
versionUpgradeOption
String
Deployment model version upgrade options:
OnceNewDefaultVersionAvailable
OnceCurrentVersionExpired
NoAutoUpgrade
capacity
integer
This represents the amount of quota you are assigning to this deployment. A value of 1 equals 1,000 Tokens per Minute (TPM)
Example request
curl -X PUT https://management.azure.com/subscriptions/00000000-0000-0000-0000-000000000000/resourceGroups/resource-group-temp/providers/Microsoft.CognitiveServices/accounts/docs-openai-test-001/deployments/text-embedding-ada-002-test-1?api-version=2023-05-01 \
  -H "Content-Type: application/json" \
  -H 'Authorization: Bearer YOUR_AUTH_TOKEN' \
  -d '{"sku":{"name":"Standard","capacity":1},"properties": {"model": {"format": "OpenAI","name": "text-embedding-ada-002","version": "2"},"versionUpgradeOption":"OnceCurrentVersionExpired"}}'
There are multiple ways to generate an authorization token. The easiest method for initial testing is to launch the Cloud Shell from the Azure portal. Then run az account get-access-token. You can use this token as your temporary authorization token for API testing.
Example response
  "id": "/subscriptions/{subscription-id}/resourceGroups/resource-group-temp/providers/Microsoft.CognitiveServices/accounts/docs-openai-test-001/deployments/text-embedding-ada-002-test-1",
  "type": "Microsoft.CognitiveServices/accounts/deployments",
  "name": "text-embedding-ada-002-test-1",
  "sku": {
    "name": "Standard",
    "capacity": 1
  "properties": {
    "model": {
      "format": "OpenAI",
      "name": "text-embedding-ada-002",
      "version": "2"
    "versionUpgradeOption": "OnceCurrentVersionExpired",
    "capabilities": {
      "embeddings": "true",
      "embeddingsMaxInputs": "1"
    "provisioningState": "Succeeded",
    "ratelimits": [
        "key": "request",
        "renewalPeriod": 10,
        "count": 2
        "key": "token",
        "renewalPeriod": 60,
        "count": 1000
  "systemData": {
    "createdBy": "docs@contoso.com",
    "createdByType": "User",
    "createdAt": "2023-06-13T00:12:38.885937Z",
    "lastModifiedBy": "docs@contoso.com",
    "lastModifiedByType": "User",
    "lastModifiedAt": "2023-06-13T02:41:04.8410965Z"
  "etag": "\"{GUID}\""
Next steps
Learn more about Azure OpenAI
Learn more about fine-tuning Azure OpenAI models