diff --git a/litellm/main.py b/litellm/main.py
index 9ab017bbbd..859486ae6f 100644
--- a/litellm/main.py
+++ b/litellm/main.py
@@ -195,7 +195,9 @@ def completion(
                 response = openai.ChatCompletion.create(
                     engine=model, messages=messages, **optional_params
                 )
-
+            if "stream" in optional_params and optional_params["stream"] == True:
+                response = CustomStreamWrapper(response, model)
+                return response
             ## LOGGING
             logging.post_call(
                 input=messages,
@@ -251,6 +253,9 @@ def completion(
                 response = openai.ChatCompletion.create(
                     model=model, messages=messages, **optional_params
                 )
+            if "stream" in optional_params and optional_params["stream"] == True:
+                response = CustomStreamWrapper(response, model)
+                return response
             ## LOGGING
             logging.post_call(
                 input=messages,
diff --git a/litellm/utils.py b/litellm/utils.py
index 8a31e36eae..3ef075a180 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -1623,6 +1623,9 @@ class CustomStreamWrapper:
         elif self.model in litellm.open_ai_text_completion_models:
             chunk = next(self.completion_stream)
             completion_obj["content"] = self.handle_openai_text_completion_chunk(chunk)
+        else: # openai chat/azure models
+            chunk = next(self.completion_stream)
+            completion_obj['content'] = chunk['choices']['delta']
         # LOGGING
         self.logging_obj(completion_obj["content"])
         # return this for all models