diff --git a/litellm/router.py b/litellm/router.py
index 940e3c90d..0688dc61f 100644
--- a/litellm/router.py
+++ b/litellm/router.py
@@ -285,7 +285,7 @@ class Router:
                 "messages": messages,
                 "functions": functions,
                 "function_call": function_call,
-                "timeout": timeout,
+                "timeout": timeout or self.timeout,
                 "temperature": temperature,
                 "top_p": top_p,
                 "n": n,
@@ -316,7 +316,7 @@ class Router:
                 future = executor.submit(
                     self.function_with_fallbacks, **kwargs, **completion_kwargs
                 )
-                response = future.result(timeout=timeout)  # type: ignore
+                response = future.result()  # type: ignore
 
             return response
         except Exception as e:
@@ -417,7 +417,7 @@ class Router:
                 "messages": messages,
                 "functions": functions,
                 "function_call": function_call,
-                "timeout": timeout,
+                "timeout": timeout or self.timeout,
                 "temperature": temperature,
                 "top_p": top_p,
                 "n": n,
@@ -442,7 +442,6 @@ class Router:
                 "original_function": self._acompletion,
             }
             kwargs["num_retries"] = kwargs.get("num_retries", self.num_retries)
-            timeout = kwargs.get("request_timeout", self.timeout)
             kwargs.setdefault("metadata", {}).update({"model_group": model})
 
             response = await self.async_function_with_fallbacks(
diff --git a/litellm/tests/test_timeout.py b/litellm/tests/test_timeout.py
index 84aa7537f..1902e1a36 100644
--- a/litellm/tests/test_timeout.py
+++ b/litellm/tests/test_timeout.py
@@ -39,6 +39,8 @@ def test_timeout():
 
 def test_hanging_request_azure():
     litellm.set_verbose = True
+    import asyncio
+
     try:
         router = litellm.Router(
             model_list=[
@@ -58,13 +60,20 @@ def test_hanging_request_azure():
         )
 
         encoded = litellm.utils.encode(model="gpt-3.5-turbo", text="blue")[0]
-        response = router.completion(
-            model="azure-gpt",
-            messages=[{"role": "user", "content": f"what color is red {uuid.uuid4()}"}],
-            logit_bias={encoded: 100},
-            timeout=0.01,
-        )
-        print(response)
+
+        async def _test():
+            response = await router.acompletion(
+                model="azure-gpt",
+                messages=[
+                    {"role": "user", "content": f"what color is red {uuid.uuid4()}"}
+                ],
+                logit_bias={encoded: 100},
+                timeout=0.01,
+            )
+            print(response)
+            return response
+
+        response = asyncio.run(_test())
 
         if response.choices[0].message.content is not None:
             pytest.fail("Got a response, expected a timeout")