diff --git a/fastrepl-proxy b/fastrepl-proxy
index c765f07b74..f2fe83e002 160000
--- a/fastrepl-proxy
+++ b/fastrepl-proxy
@@ -1 +1 @@
-Subproject commit c765f07b74f9a8cae211584ee70bad10e1a847a9
+Subproject commit f2fe83e002a7c3ddedf4e500665644adfd31b9fc
diff --git a/litellm/__pycache__/main.cpython-311.pyc b/litellm/__pycache__/main.cpython-311.pyc
index e31886e73f..7bed614a6c 100644
Binary files a/litellm/__pycache__/main.cpython-311.pyc and b/litellm/__pycache__/main.cpython-311.pyc differ
diff --git a/litellm/__pycache__/utils.cpython-311.pyc b/litellm/__pycache__/utils.cpython-311.pyc
index 1aed540c39..4f22f8da54 100644
Binary files a/litellm/__pycache__/utils.cpython-311.pyc and b/litellm/__pycache__/utils.cpython-311.pyc differ
diff --git a/litellm/main.py b/litellm/main.py
index 98033a66ed..d099c1c177 100644
--- a/litellm/main.py
+++ b/litellm/main.py
@@ -163,6 +163,7 @@ def completion(
             custom_llm_provider = model.split("/", 1)[0]
             model = model.split("/", 1)[1]
         model, custom_llm_provider = get_llm_provider(model=model, custom_llm_provider=custom_llm_provider)
+        print(f"model: {model}; llm provider: {custom_llm_provider}")
         # check if user passed in any of the OpenAI optional params
         optional_params = get_optional_params(
             functions=functions,
@@ -245,7 +246,7 @@ def completion(
                 **optional_params,
             )
             if "stream" in optional_params and optional_params["stream"] == True:
-                response = CustomStreamWrapper(response, model, logging_obj=logging)
+                response = CustomStreamWrapper(response, model, custom_llm_provider="openai", logging_obj=logging)
                 return response
             ## LOGGING
             logging.post_call(
@@ -310,7 +311,7 @@ def completion(
                 raise e
             
             if "stream" in optional_params and optional_params["stream"] == True:
-                response = CustomStreamWrapper(response, model, logging_obj=logging)
+                response = CustomStreamWrapper(response, model, custom_llm_provider="openai", logging_obj=logging)
                 return response
             ## LOGGING
             logging.post_call(
@@ -374,7 +375,7 @@ def completion(
                 **optional_params
             )
             if "stream" in optional_params and optional_params["stream"] == True:
-                response = CustomStreamWrapper(response, model, custom_llm_provider="openai", logging_obj=logging)
+                response = CustomStreamWrapper(response, model, custom_llm_provider="text-completion-openai", logging_obj=logging)
                 return response
             ## LOGGING
             logging.post_call(
@@ -446,7 +447,7 @@ def completion(
             )
             if "stream" in optional_params and optional_params["stream"] == True:
                 # don't try to access stream object,
-                response = CustomStreamWrapper(model_response, model, logging_obj=logging)
+                response = CustomStreamWrapper(model_response, model, custom_llm_provider="anthropic", logging_obj=logging)
                 return response
             response = model_response
         elif model in litellm.nlp_cloud_models or custom_llm_provider == "nlp_cloud":
@@ -493,7 +494,7 @@ def completion(
 
             if "stream" in optional_params and optional_params["stream"] == True:
                 # don't try to access stream object,
-                response = CustomStreamWrapper(model_response, model, logging_obj=logging)
+                response = CustomStreamWrapper(model_response, model, custom_llm_provider="aleph-alpha", logging_obj=logging)
                 return response
             response = model_response
         elif model in litellm.openrouter_models or custom_llm_provider == "openrouter":
diff --git a/litellm/utils.py b/litellm/utils.py
index e1d0853768..7e6b961fa3 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -2510,11 +2510,11 @@ class CustomStreamWrapper:
         model_response = ModelResponse(stream=True, model=self.model)
         try:
             # return this for all models
+            completion_obj = {"content": ""}
             if self.sent_first_chunk == False:
-                model_response.choices[0].delta.role = "assistant"
+                completion_obj["role"] = "assistant"
                 self.sent_first_chunk = True
-            completion_obj = {"content": ""} # default to role being assistant
-            if self.model in litellm.anthropic_models:
+            if self.custom_llm_provider and self.custom_llm_provider == "anthropic":
                 chunk = next(self.completion_stream)
                 completion_obj["content"] = self.handle_anthropic_chunk(chunk)
             elif self.model == "replicate" or self.custom_llm_provider == "replicate":
@@ -2539,10 +2539,10 @@ class CustomStreamWrapper:
             elif self.custom_llm_provider and self.custom_llm_provider == "vllm":
                 chunk = next(self.completion_stream)
                 completion_obj["content"] = chunk[0].outputs[0].text
-            elif self.model in litellm.aleph_alpha_models: #aleph alpha doesn't provide streaming
+            elif self.custom_llm_provider and self.custom_llm_provider == "aleph-alpha": #aleph alpha doesn't provide streaming
                 chunk = next(self.completion_stream)
                 completion_obj["content"] = self.handle_aleph_alpha_chunk(chunk)
-            elif self.model in litellm.open_ai_text_completion_models:
+            elif self.custom_llm_provider and self.custom_llm_provider == "text-completion-openai":
                 chunk = next(self.completion_stream)
                 completion_obj["content"] = self.handle_openai_text_completion_chunk(chunk)
             elif self.model in litellm.nlp_cloud_models or self.custom_llm_provider == "nlp_cloud":
@@ -2551,7 +2551,7 @@ class CustomStreamWrapper:
             elif self.model in (litellm.vertex_chat_models + litellm.vertex_code_chat_models + litellm.vertex_text_models + litellm.vertex_code_text_models):
                 chunk = next(self.completion_stream)
                 completion_obj["content"] = str(chunk)
-            elif self.model in litellm.cohere_models or self.custom_llm_provider == "cohere":
+            elif self.custom_llm_provider == "cohere":
                 chunk = next(self.completion_stream)
                 completion_obj["content"] = self.handle_cohere_chunk(chunk)
             elif self.custom_llm_provider == "bedrock":
@@ -2566,7 +2566,8 @@ class CustomStreamWrapper:
             # LOGGING
             threading.Thread(target=self.logging_obj.success_handler, args=(completion_obj,)).start()
             model_response.model = self.model
-            model_response.choices[0].delta["content"] = completion_obj["content"]
+            if len(completion_obj["content"]) > 0: # cannot set content of an OpenAI Object to be an empty string
+                model_response.choices[0].delta = Delta(**completion_obj)
             return model_response
         except StopIteration:
             raise StopIteration
diff --git a/pyproject.toml b/pyproject.toml
index ca729f02e2..812377984b 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "litellm"
-version = "0.1.683"
+version = "0.1.684"
 description = "Library to easily interface with LLM API providers"
 authors = ["BerriAI"]
 license = "MIT License"