diff --git a/litellm/llms/azure.py b/litellm/llms/azure.py
index 47ea7c0aa87..0371fc6c667 100644
--- a/litellm/llms/azure.py
+++ b/litellm/llms/azure.py
@@ -111,7 +111,7 @@ class AzureChatCompletion(BaseLLM):
         exception_mapping_worked = False
         try:
 
-            if model is None or messages is None:
+            if (model is None or messages is None) and client is None:
                 raise AzureOpenAIError(status_code=422, message=f"Missing model or messages")
         
             data = {
diff --git a/litellm/main.py b/litellm/main.py
index 850d2567239..fa5a8c868ae 100644
--- a/litellm/main.py
+++ b/litellm/main.py
@@ -494,6 +494,31 @@ def completion(
                 if k not in optional_params: # completion(top_k=3) > azure_config(top_k=3) <- allows for dynamic variables to be passed in
                     optional_params[k] = v
 
+            ### CHECK IF CLOUDFLARE AI GATEWAY ###
+            ### if so - set the model as part of the base url 
+            if "gateway.ai.cloudflare.com" in api_base and client is None: 
+                ## build base url - assume api base includes resource name
+                if not api_base.endswith("/"): 
+                    api_base += "/"
+                api_base += f"{model}"
+                
+                azure_client_params = {
+                    "api_version": api_version,
+                    "base_url": f"{api_base}",
+                    "http_client": litellm.client_session,
+                    "max_retries": max_retries,
+                    "timeout": timeout
+                }
+                if api_key is not None:
+                    azure_client_params["api_key"] = api_key
+                elif azure_ad_token is not None:
+                    azure_client_params["azure_ad_token"] = azure_ad_token
+                if acompletion is True:
+                    client = openai.AsyncAzureOpenAI(**azure_client_params)
+                else:
+                    client = openai.AzureOpenAI(**azure_client_params)
+                model = None
+
             ## COMPLETION CALL
             response = azure_chat_completions.completion(
                 model=model,
diff --git a/litellm/tests/test_completion.py b/litellm/tests/test_completion.py
index 819f08e1b54..8412d0f192a 100644
--- a/litellm/tests/test_completion.py
+++ b/litellm/tests/test_completion.py
@@ -933,7 +933,7 @@ def test_replicate_custom_prompt_dict():
     print(f"response: {response}")
     litellm.custom_prompt_dict = {} # reset 
 
-test_replicate_custom_prompt_dict() 
+# test_replicate_custom_prompt_dict() 
 
 # commenthing this out since we won't be always testing a custom replicate deployment
 # def test_completion_replicate_deployments():
@@ -1323,6 +1323,22 @@ def test_completion_anyscale_api():
 
 # test_completion_anyscale_api()
 
+def test_azure_cloudflare_api(): 
+    try: 
+        messages = [
+                {
+                    "role": "user",
+                    "content": "How do I output all files in a directory using Python?",
+                },
+            ]
+        response = completion(model="azure/gpt-turbo", messages=messages, base_url=os.getenv("CLOUDFLARE_AZURE_BASE_URL"), api_key=os.getenv("AZURE_FRANCE_API_KEY"))
+        print(f"response: {response}")
+    except Exception as e: 
+        print(e)
+        pass
+
+test_azure_cloudflare_api() 
+
 def test_completion_anyscale_2():
     try:
         # litellm.set_verbose=True