docs: add Azure AI Foundry documentation for Claude models

[Chesars]

Title

docs: add Azure AI Foundry documentation for Claude models

Relevant issues

Fixes #17066

Type

📖 Documentation

Changes

Add documentation for using Claude models (Sonnet 4.5, Haiku 4.5, Opus 4.1) deployed on Azure AI Foundry.

Azure exposes Claude using Anthropic's native /v1/messages API, so users can use the existing anthropic/ provider with their Azure endpoint - no changes to model pricing or new provider needed.

Documentation includes:

• LiteLLM SDK usage example
• Environment variables configuration
• LiteLLM Proxy config.yaml example
• Instructions for finding Azure endpoint

[vercel]

@Chesars is attempting to deploy a commit to the CLERKIEAI Team on Vercel.

A member of the Team first needs to authorize it.

[jasverma-hub]

Is this task completed or still pending? I'm trying to add the Claude model deployed on Azure AI Foundry, but it's not working for me. I've followed the documentation and added everything, yet it still doesn't work.

[Chesars]

Hi @jasverma-hub. This PR adds in the documentation an alternative method to use Anthropic models deployed in Azure, since the Azure API with Anthropic models is almost the same as the direct Anthropic API. In this PR #17202 the integration for azure with Anthropic models was added as a built-in method to use it like model=f"azure_ai/{deployment_name}", and adding to .env AZURE_AI_API_BASE and AZURE_AI_API_KEY . Are you getting any error messages from LiteLLM?

[jasverma-hub]

Hi @jasverma-hub. This PR adds in the documentation an alternative method to use Anthropic models deployed in Azure, since the Azure API with Anthropic models is almost the same as the direct Anthropic API. In this PR #17202 the integration for azure with Anthropic models was added as a built-in method to use it like model=f"azure_ai/{deployment_name}", and adding to .env AZURE_AI_API_BASE and AZURE_AI_API_KEY . Are you getting any error messages from LiteLLM?

I’ve tried adding the Claude models through both the config.yaml and the LiteLLM UI, but I’m still getting the same routing error. I’m already running the latest LiteLLM version.

This is how I currently have the models defined in my config:

• model_name: claude-sonnet-4-5
  litellm_params:
  model: azure_ai/claude-sonnet-4-5
  api_base: https://######.services.ai.azure.com/anthropic
  api_key: os.environ/OPENAI_API_KEY

• model_name: claude-opus-4-1
  litellm_params:
  model: azure_ai/claude-opus-4-1
  api_base: https://######.services.ai.azure.com/anthropic
  api_key: os.environ/OPENAI_API_KEY

• model_name: claude-opus-4-5
  litellm_params:
  model: azure_ai/claude-opus-4-5
  api_base: https://######.services.ai.azure.com/anthropic
  api_key: os.environ/OPENAI_API_KEY

[Chesars]

Hi @jasverma-hub. This PR adds in the documentation an alternative method to use Anthropic models deployed in Azure, since the Azure API with Anthropic models is almost the same as the direct Anthropic API. In this PR #17202 the integration for azure with Anthropic models was added as a built-in method to use it like model=f"azure_ai/{deployment_name}", and adding to .env AZURE_AI_API_BASE and AZURE_AI_API_KEY . Are you getting any error messages from LiteLLM?

I’ve tried adding the Claude models through both the config.yaml and the LiteLLM UI, but I’m still getting the same routing error. I’m already running the latest LiteLLM version.

This is how I currently have the models defined in my config:

• model_name: claude-sonnet-4-5

  litellm_params:

  model: azure_ai/claude-sonnet-4-5
  
  api_base: https://######.services.ai.azure.com/anthropic
  
  api_key: os.environ/OPENAI_API_KEY
  

• model_name: claude-opus-4-1

  litellm_params:

  model: azure_ai/claude-opus-4-1
  
  api_base: https://######.services.ai.azure.com/anthropic
  
  api_key: os.environ/OPENAI_API_KEY
  

• model_name: claude-opus-4-5

  litellm_params:

  model: azure_ai/claude-opus-4-5
  
  api_base: https://######.services.ai.azure.com/anthropic
  
  api_key: os.environ/OPENAI_API_KEY
  

The feature for model=azure_ai/claude-sonnet-4-5 should be available in the next release.

[jasverma-hub]

Hi @jasverma-hub. This PR adds in the documentation an alternative method to use Anthropic models deployed in Azure, since the Azure API with Anthropic models is almost the same as the direct Anthropic API. In this PR #17202 the integration for azure with Anthropic models was added as a built-in method to use it like model=f"azure_ai/{deployment_name}", and adding to .env AZURE_AI_API_BASE and AZURE_AI_API_KEY . Are you getting any error messages from LiteLLM?

I’ve tried adding the Claude models through both the config.yaml and the LiteLLM UI, but I’m still getting the same routing error. I’m already running the latest LiteLLM version.
This is how I currently have the models defined in my config:

• model_name: claude-sonnet-4-5
  litellm_params:

  model: azure_ai/claude-sonnet-4-5
  
  api_base: https://######.services.ai.azure.com/anthropic
  
  api_key: os.environ/OPENAI_API_KEY
  

• model_name: claude-opus-4-1
  litellm_params:

  model: azure_ai/claude-opus-4-1
  
  api_base: https://######.services.ai.azure.com/anthropic
  
  api_key: os.environ/OPENAI_API_KEY
  

• model_name: claude-opus-4-5
  litellm_params:

  model: azure_ai/claude-opus-4-5
  
  api_base: https://######.services.ai.azure.com/anthropic
  
  api_key: os.environ/OPENAI_API_KEY
  

The feature for model=azure_ai/claude-sonnet-4-5 should be available in the next release.

Hi @jasverma-hub. This PR adds in the documentation an alternative method to use Anthropic models deployed in Azure, since the Azure API with Anthropic models is almost the same as the direct Anthropic API. In this PR #17202 the integration for azure with Anthropic models was added as a built-in method to use it like model=f"azure_ai/{deployment_name}", and adding to .env AZURE_AI_API_BASE and AZURE_AI_API_KEY . Are you getting any error messages from LiteLLM?

How can I add this to the config.yaml? Will it work if I follow the instructions in the Anthropic documentation on LiteLLM?

[Chesars]

Yes, it will work when the updated package comes out in a release.

[jasverma-hub]

I have connected my LiteLLM to OpenWebUI, and I need to use the Claude models. Therefore, I will have to wait until the updated packages are released, as there are no other options available, right?

[Chesars]

The other option is to use the alternative method documented in this PR. https://github.com/BerriAI/litellm/pull/17104/files.

[jasverma-hub]

I tried this, but I'm still encountering this error:
azureuser@openwebui-vm:/opt/docker/litellm$ sudo nano test_anthropic.py
azureuser@openwebui-vm:/opt/docker/litellm$ python3 test_anthropic.py

Give Feedback / Get Help: https://github.com/BerriAI/litellm/issues/new
LiteLLM.Info: If you need to debug this error, use `litellm._turn_on_debug()'.

Traceback (most recent call last):
File "/opt/docker/litellm/litellm/llms/anthropic/chat/handler.py", line 440, in completion
response = client.post(
File "/opt/docker/litellm/litellm/llms/custom_httpx/http_handler.py", line 707, in post
raise e
File "/opt/docker/litellm/litellm/llms/custom_httpx/http_handler.py", line 689, in post
response.raise_for_status()
File "/usr/local/lib/python3.10/dist-packages/httpx/_models.py", line 829, in raise_for_status
raise HTTPStatusError(message, request=request, response=self)
httpx.HTTPStatusError: Client error '401 PermissionDenied' for url 'https:####'
For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/401

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/opt/docker/litellm/litellm/main.py", line 2051, in completion
response = anthropic_chat_completions.completion(
File "/opt/docker/litellm/litellm/llms/anthropic/chat/handler.py", line 455, in completion
raise AnthropicError(
litellm.llms.anthropic.common_utils.AnthropicError: {"error":{"code":"401","message":"Access denied due to invalid subscription key or wrong API endpoint. Make sure to provide a valid key for an active subscription and use a correct regional API endpoint for your resource."}}

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/opt/docker/litellm/test_anthropic.py", line 3, in
response = completion(
File "/opt/docker/litellm/litellm/utils.py", line 1285, in wrapper
raise e
File "/opt/docker/litellm/litellm/utils.py", line 1163, in wrapper
result = original_function(*args, **kwargs)
File "/opt/docker/litellm/litellm/main.py", line 3305, in completion
raise exception_type(
File "/opt/docker/litellm/litellm/litellm_core_utils/exception_mapping_utils.py", line 2271, in exception_type
raise e
File "/opt/docker/litellm/litellm/litellm_core_utils/exception_mapping_utils.py", line 571, in exception_type
raise AuthenticationError(
litellm.exceptions.AuthenticationError: litellm.AuthenticationError: AnthropicException - {"error":{"code":"401","message":"Access denied due to invalid subscription key or wrong API endpoint. Make sure to provide a valid key for an active subscription and use a correct regional API endpoint for your resource."}}
azureuser@openwebui-vm:/opt/docker/litellm$

I am using the exact API BASE and API KEY from Azure still encountering this.

[Chesars]

Thanks for reporting it, we'll look into that. So yes, we'll have to wait for the next release.

[jasverma-hub]

Okay, fine. One last thing: is the problem with Claude models or Claude models hosted by Azure?