[Fixes] A2a Gateway - ensure azure foundry agents work (#17943)

* add agents  v2 fixes azure

* fix auth

* get_azure_ad_token fix

* docs foundry

docs/my-website/docs/providers/azure_ai_agents.md

@@ -9,7 +9,47 @@ Call Azure AI Foundry Agents in the OpenAI Request/Response format.
 |----------|---------|
 | Description | Azure AI Foundry Agents provides hosted agent runtimes that can execute agentic workflows with foundation models, tools, and code interpreters. |
 | Provider Route on LiteLLM | `azure_ai/agents/{AGENT_ID}` |
-| Provider Doc | [Azure AI Foundry Agents ↗](https://learn.microsoft.com/en-us/rest/api/aifoundry/aiagents/create-thread-and-run/create-thread-and-run) |
+| Provider Doc | [Azure AI Foundry Agents ↗](https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart) |
+
+## Authentication
+
+Azure AI Foundry Agents require **Azure AD authentication** (not API keys). You can authenticate using:
+
+### Option 1: Service Principal (Recommended for Production)
+
+Set these environment variables:
+
+```bash
+export AZURE_TENANT_ID="your-tenant-id"
+export AZURE_CLIENT_ID="your-client-id"
+export AZURE_CLIENT_SECRET="your-client-secret"
+```
+
+LiteLLM will automatically obtain an Azure AD token using these credentials.
+
+### Option 2: Azure AD Token (Manual)
+
+Pass a token directly via `api_key`:
+
+```bash
+# Get token via Azure CLI
+az account get-access-token --resource "https://ai.azure.com" --query accessToken -o tsv
+```
+
+### Required Azure Role
+
+Your Service Principal or user must have the **Azure AI Developer** or **Azure AI User** role on your Azure AI Foundry project.
+
+To assign via Azure CLI:
+```bash
+az role assignment create \
+  --assignee-object-id "<service-principal-object-id>" \
+  --assignee-principal-type "ServicePrincipal" \
+  --role "Azure AI Developer" \
+  --scope "/subscriptions/<sub>/resourceGroups/<rg>/providers/Microsoft.CognitiveServices/accounts/<resource>"
+```
+
+Or add via **Azure AI Foundry Portal** → Your Project → **Project users** → **+ New user**.
 
 ## Quick Start
 
@@ -34,6 +74,7 @@ You can find the Agent ID in your Azure AI Foundry portal under Agents.
 import litellm
 
 # Make a completion request to your Azure AI Foundry Agent
+# Uses AZURE_TENANT_ID, AZURE_CLIENT_ID, AZURE_CLIENT_SECRET env vars for auth
 response = litellm.completion(
     model="azure_ai/agents/asst_abc123",
     messages=[
@@ -42,8 +83,7 @@ response = litellm.completion(
             "content": "Explain machine learning in simple terms"
         }
     ],
-    api_base="https://your-project.services.ai.azure.com",
-    api_key="your-api-key",
+    api_base="https://your-resource.services.ai.azure.com/api/projects/your-project",
 )
 
 print(response.choices[0].message.content)
@@ -62,8 +102,7 @@ response = await litellm.acompletion(
             "content": "What are the key principles of software architecture?"
         }
     ],
-    api_base="https://your-project.services.ai.azure.com",
-    api_key="your-api-key",
+    api_base="https://your-resource.services.ai.azure.com/api/projects/your-project",
     stream=True,
 )
 
@@ -84,14 +123,18 @@ model_list:
   - model_name: azure-agent-1
     litellm_params:
       model: azure_ai/agents/asst_abc123
-      api_base: https://your-project.services.ai.azure.com
-      api_key: os.environ/AZURE_API_KEY
+      api_base: https://your-resource.services.ai.azure.com/api/projects/your-project
+      # Service Principal auth (recommended)
+      tenant_id: os.environ/AZURE_TENANT_ID
+      client_id: os.environ/AZURE_CLIENT_ID
+      client_secret: os.environ/AZURE_CLIENT_SECRET
 
   - model_name: azure-agent-math-tutor
     litellm_params:
       model: azure_ai/agents/asst_def456
-      api_base: https://your-project.services.ai.azure.com
-      api_key: os.environ/AZURE_API_KEY
+      api_base: https://your-resource.services.ai.azure.com/api/projects/your-project
+      # Or pass Azure AD token directly
+      api_key: os.environ/AZURE_AD_TOKEN
 ```
 
 </TabItem>
@@ -196,16 +239,16 @@ for chunk in stream:
 
 ## Environment Variables
 
-You can set the following environment variables to configure Azure AI Foundry Agents:
-
 | Variable | Description |
 |----------|-------------|
-| `AZURE_API_BASE` | The Azure AI Foundry project endpoint (e.g., `https://your-project.services.ai.azure.com`) |
-| `AZURE_API_KEY` | Your Azure AI Foundry API key |
+| `AZURE_TENANT_ID` | Azure AD tenant ID for Service Principal auth |
+| `AZURE_CLIENT_ID` | Application (client) ID of your Service Principal |
+| `AZURE_CLIENT_SECRET` | Client secret for your Service Principal |
 
 ```bash
-export AZURE_API_BASE="https://your-project.services.ai.azure.com"
-export AZURE_API_KEY="your-api-key"
+export AZURE_TENANT_ID="your-tenant-id"
+export AZURE_CLIENT_ID="your-client-id"
+export AZURE_CLIENT_SECRET="your-client-secret"
 ```
 
 ## Conversation Continuity (Thread Management)
@@ -219,8 +262,7 @@ import litellm
 response1 = await litellm.acompletion(
     model="azure_ai/agents/asst_abc123",
     messages=[{"role": "user", "content": "My name is Alice"}],
-    api_base="https://your-project.services.ai.azure.com",
-    api_key="your-api-key",
+    api_base="https://your-resource.services.ai.azure.com/api/projects/your-project",
 )
 
 # Get the thread_id from the response
@@ -230,8 +272,7 @@ thread_id = response1._hidden_params.get("thread_id")
 response2 = await litellm.acompletion(
     model="azure_ai/agents/asst_abc123",
     messages=[{"role": "user", "content": "What's my name?"}],
-    api_base="https://your-project.services.ai.azure.com",
-    api_key="your-api-key",
+    api_base="https://your-resource.services.ai.azure.com/api/projects/your-project",
     thread_id=thread_id,  # Pass the thread_id to continue conversation
 )
 
@@ -256,8 +297,7 @@ response = litellm.completion(
             "content": "Analyze this data and provide insights",
         }
     ],
-    api_base="https://your-project.services.ai.azure.com",
-    api_key="your-api-key",
+    api_base="https://your-resource.services.ai.azure.com/api/projects/your-project",
     thread_id="thread_abc123",  # Optional: Continue existing conversation
     instructions="Be concise and focus on key insights",  # Optional: Override agent instructions
 )
@@ -271,8 +311,10 @@ model_list:
   - model_name: azure-agent-analyst
     litellm_params:
       model: azure_ai/agents/asst_abc123
-      api_base: https://your-project.services.ai.azure.com
-      api_key: os.environ/AZURE_API_KEY
+      api_base: https://your-resource.services.ai.azure.com/api/projects/your-project
+      tenant_id: os.environ/AZURE_TENANT_ID
+      client_id: os.environ/AZURE_CLIENT_ID
+      client_secret: os.environ/AZURE_CLIENT_SECRET
       instructions: "Be concise and focus on key insights"
 ```
 

litellm/llms/azure/common_utils.py

@@ -294,20 +294,18 @@ def get_azure_ad_token(
         Azure AD token as string if successful, None otherwise
     """
     # Extract parameters
+    # Use `or` instead of default parameter to handle cases where key exists but value is None
     azure_ad_token_provider = litellm_params.get("azure_ad_token_provider")
-    azure_ad_token = litellm_params.get("azure_ad_token", None) or get_secret_str(
+    azure_ad_token = litellm_params.get("azure_ad_token") or get_secret_str(
         "AZURE_AD_TOKEN"
     )
-    tenant_id = litellm_params.get("tenant_id", os.getenv("AZURE_TENANT_ID"))
-    client_id = litellm_params.get("client_id", os.getenv("AZURE_CLIENT_ID"))
-    client_secret = litellm_params.get(
-        "client_secret", os.getenv("AZURE_CLIENT_SECRET")
-    )
-    azure_username = litellm_params.get("azure_username", os.getenv("AZURE_USERNAME"))
-    azure_password = litellm_params.get("azure_password", os.getenv("AZURE_PASSWORD"))
-    scope = litellm_params.get(
-        "azure_scope",
-        os.getenv("AZURE_SCOPE", "https://cognitiveservices.azure.com/.default"),
+    tenant_id = litellm_params.get("tenant_id") or os.getenv("AZURE_TENANT_ID")
+    client_id = litellm_params.get("client_id") or os.getenv("AZURE_CLIENT_ID")
+    client_secret = litellm_params.get("client_secret") or os.getenv("AZURE_CLIENT_SECRET")
+    azure_username = litellm_params.get("azure_username") or os.getenv("AZURE_USERNAME")
+    azure_password = litellm_params.get("azure_password") or os.getenv("AZURE_PASSWORD")
+    scope = litellm_params.get("azure_scope") or os.getenv(
+        "AZURE_SCOPE", "https://cognitiveservices.azure.com/.default"
     )
     if scope is None:
         scope = "https://cognitiveservices.azure.com/.default"

litellm/llms/azure_ai/agents/handler.py

@@ -1,5 +1,5 @@
 """
-Handler for Azure AI Agent Service API.
+Handler for Azure Foundry Agent Service API.
 
 This handler executes the multi-step agent flow:
 1. Create thread (or use existing)
@@ -8,8 +8,14 @@ This handler executes the multi-step agent flow:
 4. Retrieve the assistant's response messages
 
 Model format: azure_ai/agents/<agent_id>
+API Base format: https://<AIFoundryResourceName>.services.ai.azure.com/api/projects/<ProjectName>
+
+Authentication: Uses Azure AD Bearer tokens (not API keys)
+  Get token via: az account get-access-token --resource 'https://ai.azure.com'
 
 Supports both polling-based and native streaming (SSE) modes.
+
+See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
 """
 
 import asyncio
@@ -60,24 +66,27 @@ class AzureAIAgentsHandler:
     # -------------------------------------------------------------------------
     # URL Builders
     # -------------------------------------------------------------------------
+    # Azure Foundry Agents API uses /assistants, /threads, etc. directly
+    # See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
+    # -------------------------------------------------------------------------
     def _build_thread_url(self, api_base: str, api_version: str) -> str:
-        return f"{api_base}/openai/threads?api-version={api_version}"
+        return f"{api_base}/threads?api-version={api_version}"
 
     def _build_messages_url(self, api_base: str, thread_id: str, api_version: str) -> str:
-        return f"{api_base}/openai/threads/{thread_id}/messages?api-version={api_version}"
+        return f"{api_base}/threads/{thread_id}/messages?api-version={api_version}"
 
     def _build_runs_url(self, api_base: str, thread_id: str, api_version: str) -> str:
-        return f"{api_base}/openai/threads/{thread_id}/runs?api-version={api_version}"
+        return f"{api_base}/threads/{thread_id}/runs?api-version={api_version}"
 
     def _build_run_status_url(self, api_base: str, thread_id: str, run_id: str, api_version: str) -> str:
-        return f"{api_base}/openai/threads/{thread_id}/runs/{run_id}?api-version={api_version}"
+        return f"{api_base}/threads/{thread_id}/runs/{run_id}?api-version={api_version}"
 
     def _build_list_messages_url(self, api_base: str, thread_id: str, api_version: str) -> str:
-        return f"{api_base}/openai/threads/{thread_id}/messages?api-version={api_version}"
+        return f"{api_base}/threads/{thread_id}/messages?api-version={api_version}"
 
     def _build_create_thread_and_run_url(self, api_base: str, api_version: str) -> str:
         """URL for the create-thread-and-run endpoint (supports streaming)."""
-        return f"{api_base}/openai/threads/runs?api-version={api_version}"
+        return f"{api_base}/threads/runs?api-version={api_version}"
 
     # -------------------------------------------------------------------------
     # Response Helpers
@@ -140,12 +149,21 @@ class AzureAIAgentsHandler:
         optional_params: dict,
         headers: Optional[dict],
     ) -> tuple:
-        """Prepare common parameters for completion."""
+        """Prepare common parameters for completion.
+        
+        Azure Foundry Agents API uses Bearer token authentication:
+        - Authorization: Bearer <token> (Azure AD token from 'az account get-access-token --resource https://ai.azure.com')
+        
+        See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
+        """
         if headers is None:
             headers = {}
         headers["Content-Type"] = "application/json"
+        
+        # Azure Foundry Agents uses Bearer token authentication
+        # The api_key here is expected to be an Azure AD token
         if api_key:
-            headers["api-key"] = api_key
+            headers["Authorization"] = f"Bearer {api_key}"
 
         api_version = optional_params.get("api_version", self.config.DEFAULT_API_VERSION)
         agent_id = self.config._get_agent_id(model, optional_params)

litellm/llms/azure_ai/agents/transformation.py

@@ -1,17 +1,24 @@
 """
-Transformation for Azure AI Agent Service API.
+Transformation for Azure Foundry Agent Service API.
 
-Azure AI Agent Service provides an Assistants-like API for running agents.
+Azure Foundry Agent Service provides an Assistants-like API for running agents.
 This follows the OpenAI Assistants pattern: create thread -> add messages -> create/poll run.
 
 Model format: azure_ai/agents/<agent_id>
 
+API Base format: https://<AIFoundryResourceName>.services.ai.azure.com/api/projects/<ProjectName>
+
+Authentication: Uses Azure AD Bearer tokens (not API keys)
+  Get token via: az account get-access-token --resource 'https://ai.azure.com'
+
 The API uses these endpoints:
-- POST /openai/threads - Create a thread
-- POST /openai/threads/{thread_id}/messages - Add message to thread
-- POST /openai/threads/{thread_id}/runs - Create a run
-- GET /openai/threads/{thread_id}/runs/{run_id} - Poll run status
-- GET /openai/threads/{thread_id}/messages - List messages in thread
+- POST /threads - Create a thread
+- POST /threads/{thread_id}/messages - Add message to thread
+- POST /threads/{thread_id}/runs - Create a run
+- GET /threads/{thread_id}/runs/{run_id} - Poll run status
+- GET /threads/{thread_id}/messages - List messages in thread
+
+See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
 """
 
 from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Union
@@ -59,8 +66,10 @@ class AzureAIAgentsConfig(BaseConfig):
     4. Retrieve the assistant's response messages
     """
 
-    # Default API version for Azure AI Agent Service
-    DEFAULT_API_VERSION = "2024-07-01-preview"
+    # Default API version for Azure Foundry Agent Service
+    # GA version: 2025-05-01, Preview: 2025-05-15-preview
+    # See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
+    DEFAULT_API_VERSION = "2025-05-01"
     
     # Polling configuration
     MAX_POLL_ATTEMPTS = 60
@@ -236,13 +245,19 @@ class AzureAIAgentsConfig(BaseConfig):
         api_base: Optional[str] = None,
     ) -> dict:
         """
-        Validate and set up environment for Azure Agents requests.
+        Validate and set up environment for Azure Foundry Agents requests.
+        
+        Azure Foundry Agents uses Bearer token authentication with Azure AD tokens.
+        Get token via: az account get-access-token --resource 'https://ai.azure.com'
+        
+        See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
         """
         headers["Content-Type"] = "application/json"
 
-        # Add API key if provided
+        # Azure Foundry Agents uses Bearer token authentication
+        # The api_key here is expected to be an Azure AD token
         if api_key:
-            headers["api-key"] = api_key
+            headers["Authorization"] = f"Bearer {api_key}"
 
         return headers
 
@@ -310,15 +325,38 @@ class AzureAIAgentsConfig(BaseConfig):
         headers: Optional[dict] = None,
     ) -> Any:
         """
-        Dispatch method for Azure AI Agents completion.
+        Dispatch method for Azure Foundry Agents completion.
         
         Routes to sync or async completion based on acompletion flag.
         Supports native streaming via SSE when stream=True and acompletion=True.
+        
+        Authentication: Uses Azure AD Bearer tokens.
+        - Pass api_key directly as an Azure AD token
+        - Or set up Azure AD credentials via environment variables for automatic token retrieval:
+          - AZURE_TENANT_ID, AZURE_CLIENT_ID, AZURE_CLIENT_SECRET (Service Principal)
+        
+        See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
         """
+        from litellm.llms.azure.common_utils import get_azure_ad_token
         from litellm.llms.azure_ai.agents.handler import azure_ai_agents_handler
+        from litellm.types.router import GenericLiteLLMParams
 
+        # If no api_key is provided, try to get Azure AD token
         if api_key is None:
-            raise ValueError("api_key is required for Azure AI Agents")
+            # Try to get Azure AD token using the existing Azure auth mechanisms
+            # This uses the scope for Azure AI (ai.azure.com) instead of cognitive services
+            # Create a GenericLiteLLMParams with the scope override for Azure Foundry Agents
+            azure_auth_params = dict(litellm_params) if litellm_params else {}
+            azure_auth_params["azure_scope"] = "https://ai.azure.com/.default"
+            api_key = get_azure_ad_token(GenericLiteLLMParams(**azure_auth_params))
+            
+        if api_key is None:
+            raise ValueError(
+                "api_key (Azure AD token) is required for Azure Foundry Agents. "
+                "Either pass api_key directly, or set AZURE_TENANT_ID, AZURE_CLIENT_ID, "
+                "and AZURE_CLIENT_SECRET environment variables for Service Principal auth. "
+                "Manual token: az account get-access-token --resource 'https://ai.azure.com'"
+            )
         if acompletion:
             if stream:
                 # Native async streaming via SSE - return the async generator directly

litellm/proxy/public_endpoints/agent_create_fields.json

@@ -93,8 +93,8 @@
       {
         "key": "api_base",
         "label": "Azure AI API Base",
-        "placeholder": "https://your-project.services.ai.azure.com",
-        "tooltip": "The base URL for your Azure AI Foundry project endpoint",
+        "placeholder": "https://your-resource.services.ai.azure.com/api/projects/your-project",
+        "tooltip": "The base URL for your Azure AI Foundry project endpoint (e.g., https://your-resource.services.ai.azure.com/api/projects/your-project)",
         "required": true,
         "field_type": "text",
         "default_value": null,
@@ -102,10 +102,40 @@
       },
       {
         "key": "api_key",
-        "label": "Azure AI API Key",
+        "label": "Azure AD Token",
         "placeholder": null,
-        "tooltip": "API key for authenticating with your Azure AI Foundry project",
-        "required": true,
+        "tooltip": "Azure AD Bearer token for authentication. Optional if using Service Principal credentials below. Get via: az account get-access-token --resource https://ai.azure.com",
+        "required": false,
+        "field_type": "password",
+        "default_value": null,
+        "include_in_litellm_params": true
+      },
+      {
+        "key": "tenant_id",
+        "label": "Azure Tenant ID",
+        "placeholder": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
+        "tooltip": "Azure AD Tenant ID for Service Principal authentication. Find in Azure Portal > Azure Active Directory > Overview",
+        "required": false,
+        "field_type": "text",
+        "default_value": null,
+        "include_in_litellm_params": true
+      },
+      {
+        "key": "client_id",
+        "label": "Azure Client ID",
+        "placeholder": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
+        "tooltip": "Application (client) ID of your Service Principal. Find in Azure Portal > App registrations > your app",
+        "required": false,
+        "field_type": "text",
+        "default_value": null,
+        "include_in_litellm_params": true
+      },
+      {
+        "key": "client_secret",
+        "label": "Azure Client Secret",
+        "placeholder": null,
+        "tooltip": "Client secret for your Service Principal. Create in Azure Portal > App registrations > your app > Certificates & secrets",
+        "required": false,
         "field_type": "password",
         "default_value": null,
         "include_in_litellm_params": true

tests/llm_translation/test_azure_agents.py

@@ -1,9 +1,9 @@
 """
-Tests for Azure AI Agent Service integration.
+Tests for Azure Foundry Agent Service integration.
 
-These tests require an Azure AI Agent Service endpoint and a pre-configured agent.
+These tests require an Azure Foundry Agent Service endpoint and a pre-configured agent.
 
-The Azure AI Agent Service uses the Assistants API pattern:
+The Azure Foundry Agent Service uses the Assistants API pattern:
 1. Create a thread
 2. Add messages to the thread
 3. Create and poll a run
@@ -11,9 +11,16 @@ The Azure AI Agent Service uses the Assistants API pattern:
 
 Model format: azure_ai/agents/<agent_id>
 
+API Base format: https://<AIFoundryResourceName>.services.ai.azure.com/api/projects/<ProjectName>
+
+Authentication: Uses Azure AD Bearer tokens (not API keys)
+  Get token via: az account get-access-token --resource 'https://ai.azure.com'
+
 Example environment variables:
-  AZURE_AI_API_BASE=https://your-project.services.ai.azure.com
-  AZURE_AI_API_KEY=your-api-key
+  AZURE_AGENTS_API_BASE=https://litellm-ci-cd-prod.services.ai.azure.com/api/projects/litellm-ci-cd
+  AZURE_AGENTS_API_KEY=<Azure AD Bearer token>
+
+See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
 """
 
 import os
@@ -29,13 +36,15 @@ import litellm
 @pytest.mark.asyncio
 async def test_azure_ai_agents_acompletion_non_streaming():
     """
-    Test non-streaming acompletion call to Azure AI Agent Service.
+    Test non-streaming acompletion call to Azure Foundry Agent Service.
     Uses the multi-step flow: create thread -> add messages -> create/poll run -> get messages
     """
-    api_base = os.environ.get("AZURE_API_BASE")
-    api_key = os.environ.get("AZURE_API_KEY")
-    agent_id = "asst_shNRIVxMPuvSRVWP5WvVe4jE"
+    api_base = os.environ.get("AZURE_AGENTS_API_BASE")
+    api_key = os.environ.get("AZURE_AGENTS_API_KEY")
+    agent_id = os.environ.get("AZURE_AGENTS_AGENT_ID", "asst_hbnoK9BOCcHhC3lC4MDroVGG")
 
+    if not api_base or not api_key:
+        pytest.skip("AZURE_AGENTS_API_BASE and AZURE_AGENTS_API_KEY environment variables required")
 
     response = await litellm.acompletion(
         model=f"azure_ai/agents/{agent_id}",
@@ -62,12 +71,15 @@ async def test_azure_ai_agents_acompletion_non_streaming():
 @pytest.mark.asyncio
 async def test_azure_ai_agents_acompletion_streaming():
     """
-    Test native streaming acompletion call to Azure AI Agent Service.
+    Test native streaming acompletion call to Azure Foundry Agent Service.
     Uses the create-thread-and-run endpoint with stream=True for SSE streaming.
     """
-    api_base = os.environ.get("AZURE_API_BASE")
-    api_key = os.environ.get("AZURE_API_KEY")
-    agent_id = os.environ.get("AZURE_AGENTS_AGENT_ID", "asst_shNRIVxMPuvSRVWP5WvVe4jE")
+    api_base = os.environ.get("AZURE_AGENTS_API_BASE")
+    api_key = os.environ.get("AZURE_AGENTS_API_KEY")
+    agent_id = os.environ.get("AZURE_AGENTS_AGENT_ID", "asst_hbnoK9BOCcHhC3lC4MDroVGG")
+
+    if not api_base or not api_key:
+        pytest.skip("AZURE_AGENTS_API_BASE and AZURE_AGENTS_API_KEY environment variables required")
 
     response = await litellm.acompletion(
         model=f"azure_ai/agents/{agent_id}",
@@ -223,7 +235,7 @@ def test_azure_ai_agents_config_transform_request():
     assert request["messages"][0]["role"] == "system"
     assert request["messages"][1]["role"] == "user"
     assert "api_version" in request
-    assert request["api_version"] == "2024-07-01-preview"
+    assert request["api_version"] == "2025-05-01"
 
 
 def test_azure_ai_agents_provider_detection():
@@ -243,7 +255,9 @@ def test_azure_ai_agents_provider_detection():
 
 def test_azure_ai_agents_validate_environment():
     """
-    Test that headers are correctly set up.
+    Test that headers are correctly set up with Bearer token authentication.
+    
+    Azure Foundry Agents uses Bearer token authentication (Azure AD tokens).
     """
     from litellm.llms.azure_ai.agents.transformation import AzureAIAgentsConfig
 
@@ -255,41 +269,44 @@ def test_azure_ai_agents_validate_environment():
         messages=[],
         optional_params={},
         litellm_params={},
-        api_key="test-api-key",
-        api_base="https://test.services.ai.azure.com",
+        api_key="test-azure-ad-token",
+        api_base="https://test.services.ai.azure.com/api/projects/test-project",
     )
 
     assert headers["Content-Type"] == "application/json"
-    assert headers["api-key"] == "test-api-key"
+    assert headers["Authorization"] == "Bearer test-azure-ad-token"
 
 
 def test_azure_ai_agents_handler_url_builders():
     """
     Test the URL building methods in the handler.
+    
+    Azure Foundry Agents API uses direct paths without /openai/ prefix.
+    See: https://learn.microsoft.com/en-us/azure/ai-foundry/agents/quickstart
     """
     from litellm.llms.azure_ai.agents.handler import AzureAIAgentsHandler
 
     handler = AzureAIAgentsHandler()
-    api_base = "https://test.services.ai.azure.com"
-    api_version = "2024-07-01-preview"
+    api_base = "https://test.services.ai.azure.com/api/projects/test-project"
+    api_version = "2025-05-01"
     thread_id = "thread_abc123"
     run_id = "run_xyz789"
 
-    # Test thread URL - uses /openai/ prefix
+    # Test thread URL - direct path without /openai/ prefix
     thread_url = handler._build_thread_url(api_base, api_version)
-    assert thread_url == f"{api_base}/openai/threads?api-version={api_version}"
+    assert thread_url == f"{api_base}/threads?api-version={api_version}"
 
     # Test messages URL
     messages_url = handler._build_messages_url(api_base, thread_id, api_version)
-    assert messages_url == f"{api_base}/openai/threads/{thread_id}/messages?api-version={api_version}"
+    assert messages_url == f"{api_base}/threads/{thread_id}/messages?api-version={api_version}"
 
     # Test runs URL
     runs_url = handler._build_runs_url(api_base, thread_id, api_version)
-    assert runs_url == f"{api_base}/openai/threads/{thread_id}/runs?api-version={api_version}"
+    assert runs_url == f"{api_base}/threads/{thread_id}/runs?api-version={api_version}"
 
     # Test run status URL
     status_url = handler._build_run_status_url(api_base, thread_id, run_id, api_version)
-    assert status_url == f"{api_base}/openai/threads/{thread_id}/runs/{run_id}?api-version={api_version}"
+    assert status_url == f"{api_base}/threads/{thread_id}/runs/{run_id}?api-version={api_version}"
 
 
 def test_azure_ai_agents_extract_content_from_messages():
@@ -340,12 +357,12 @@ async def test_azure_ai_agents_conversation_continuity():
     """
     Test that thread_id can be used for conversation continuity.
     """
-    api_base = os.environ.get("AZURE_AI_API_BASE")
-    api_key = os.environ.get("AZURE_AI_API_KEY")
-    agent_id = os.environ.get("AZURE_AI_AGENTS_AGENT_ID", "asst_shNRIVxMPuvSRVWP5WvVe4jE")
+    api_base = os.environ.get("AZURE_AGENTS_API_BASE")
+    api_key = os.environ.get("AZURE_AGENTS_API_KEY")
+    agent_id = os.environ.get("AZURE_AGENTS_AGENT_ID", "asst_hbnoK9BOCcHhC3lC4MDroVGG")
 
     if not api_base or not api_key:
-        pytest.skip("AZURE_AI_API_BASE and AZURE_AI_API_KEY environment variables required")
+        pytest.skip("AZURE_AGENTS_API_BASE and AZURE_AGENTS_API_KEY environment variables required")
 
     try:
         # First message