diff --git a/litellm/tests/test_completion.py b/litellm/tests/test_completion.py
index bd0301f204..de79c97afa 100644
--- a/litellm/tests/test_completion.py
+++ b/litellm/tests/test_completion.py
@@ -544,13 +544,13 @@ def hf_test_completion_tgi():
 def test_completion_openai():
     try:
         litellm.set_verbose = True
+        litellm.drop_params = True
         print(f"api key: {os.environ['OPENAI_API_KEY']}")
         litellm.api_key = os.environ["OPENAI_API_KEY"]
         response = completion(
             model="gpt-3.5-turbo",
-            messages=messages,
+            messages=[{"role": "user", "content": "Hey"}],
             max_tokens=10,
-            request_timeout=1,
             metadata={"hi": "bye"},
         )
         print("This is the response object\n", response)
@@ -565,6 +565,7 @@ def test_completion_openai():
         assert len(response_str) > 1
 
         litellm.api_key = None
+        raise Exception("it works!")
     except Timeout as e:
         pass
     except Exception as e:
diff --git a/litellm/utils.py b/litellm/utils.py
index 5ccb85ef05..fdca57e51f 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -2348,7 +2348,9 @@ def client(original_function):
                     elif user_max_tokens + input_tokens > max_output_tokens:
                         user_max_tokens = max_output_tokens - input_tokens
                     print_verbose(f"user_max_tokens: {user_max_tokens}")
-                    kwargs["max_tokens"] = user_max_tokens
+                    kwargs["max_tokens"] = int(
+                        round(user_max_tokens)
+                    )  # make sure max tokens is always an int
                 except Exception as e:
                     print_verbose(f"Error while checking max token limit: {str(e)}")
             # MODEL CALL