diff --git a/src/openai/lib/azure.py b/src/openai/lib/azure.py
index ea7bd20d99..589679b1b2 100644
--- a/src/openai/lib/azure.py
+++ b/src/openai/lib/azure.py
@@ -216,11 +216,7 @@ def __init__(
                 raise ValueError(
                     "Must provide one of the `base_url` or `azure_endpoint` arguments, or the `AZURE_OPENAI_ENDPOINT` environment variable"
                 )
-
-            if azure_deployment is not None:
-                base_url = f"{azure_endpoint.rstrip('/')}/openai/deployments/{azure_deployment}"
-            else:
-                base_url = f"{azure_endpoint.rstrip('/')}/openai"
+            base_url = f"{azure_endpoint.rstrip('/')}/openai"
         else:
             if azure_endpoint is not None:
                 raise ValueError("base_url and azure_endpoint are mutually exclusive")
@@ -229,6 +225,7 @@ def __init__(
             # define a sentinel value to avoid any typing issues
             api_key = API_KEY_SENTINEL
 
+        self._azure_deployment = azure_deployment
         super().__init__(
             api_key=api_key,
             organization=organization,
@@ -337,10 +334,14 @@ def _configure_realtime(self, model: str, extra_query: Query) -> tuple[httpx.URL
             "api-version": self._api_version,
             "deployment": self._azure_deployment or model,
         }
+        if self._azure_deployment:
+            query["deployment"] = self._azure_deployment
+        else:
+            query["deployment"] = model
         if self.api_key != "<missing API key>":
             auth_headers = {"api-key": self.api_key}
         else:
-            token = self._get_azure_ad_token()
+            token = await self._get_azure_ad_token()
             if token:
                 auth_headers = {"Authorization": f"Bearer {token}"}
 
@@ -491,10 +492,7 @@ def __init__(
                     "Must provide one of the `base_url` or `azure_endpoint` arguments, or the `AZURE_OPENAI_ENDPOINT` environment variable"
                 )
 
-            if azure_deployment is not None:
-                base_url = f"{azure_endpoint.rstrip('/')}/openai/deployments/{azure_deployment}"
-            else:
-                base_url = f"{azure_endpoint.rstrip('/')}/openai"
+            base_url = f"{azure_endpoint.rstrip('/')}/openai"
         else:
             if azure_endpoint is not None:
                 raise ValueError("base_url and azure_endpoint are mutually exclusive")
@@ -613,6 +611,10 @@ async def _configure_realtime(self, model: str, extra_query: Query) -> tuple[htt
             "api-version": self._api_version,
             "deployment": self._azure_deployment or model,
         }
+        if self._azure_deployment:
+            query["deployment"] = self._azure_deployment
+        else:
+            query["deployment"] = model
         if self.api_key != "<missing API key>":
             auth_headers = {"api-key": self.api_key}
         else:
diff --git a/src/openai/resources/beta/realtime/realtime.py b/src/openai/resources/beta/realtime/realtime.py
index 76e57f8cb7..5e1278c7d0 100644
--- a/src/openai/resources/beta/realtime/realtime.py
+++ b/src/openai/resources/beta/realtime/realtime.py
@@ -362,15 +362,14 @@ async def __aenter__(self) -> AsyncRealtimeConnection:
         extra_query = self.__extra_query
         auth_headers = self.__client.auth_headers
         if is_async_azure_client(self.__client):
-            url, auth_headers = await self.__client._configure_realtime(self.__model, extra_query)
-        else:
-            url = self._prepare_url().copy_with(
-                params={
-                    **self.__client.base_url.params,
-                    "model": self.__model,
-                    **extra_query,
-                },
-            )
+            extra_query, auth_headers = await self.__client._configure_realtime(self.__model, extra_query)
+
+        url = self._prepare_url().copy_with(
+            params={
+                "model": self.__model,
+                **extra_query,
+            },
+        )
         log.debug("Connecting to %s", url)
         if self.__websocket_connection_options:
             log.debug("Connection options: %s", self.__websocket_connection_options)
@@ -546,15 +545,14 @@ def __enter__(self) -> RealtimeConnection:
         extra_query = self.__extra_query
         auth_headers = self.__client.auth_headers
         if is_azure_client(self.__client):
-            url, auth_headers = self.__client._configure_realtime(self.__model, extra_query)
-        else:
-            url = self._prepare_url().copy_with(
-                params={
-                    **self.__client.base_url.params,
-                    "model": self.__model,
-                    **extra_query,
-                },
-            )
+            extra_query, auth_headers = self.__client._configure_realtime(self.__model, extra_query)
+
+        url = self._prepare_url().copy_with(
+            params={
+                "model": self.__model,
+                **extra_query,
+            },
+        )
         log.debug("Connecting to %s", url)
         if self.__websocket_connection_options:
             log.debug("Connection options: %s", self.__websocket_connection_options)