diff --git a/tests/llm_translation/test_azure_o_series.py b/tests/llm_translation/test_azure_o_series.py
index 1a2d672af71..ce7e614cbe2 100644
--- a/tests/llm_translation/test_azure_o_series.py
+++ b/tests/llm_translation/test_azure_o_series.py
@@ -159,15 +159,23 @@ def test_azure_o_series_routing():
 def test_openai_o_series_max_retries_0(mock_get_openai_client):
     import litellm
 
+    mock_get_openai_client.return_value.chat.completions.with_raw_response.create.return_value.headers = {}
+    mock_get_openai_client.return_value.chat.completions.with_raw_response.create.return_value.parse.return_value = (
+        ModelResponse(choices=[{"message": {"role": "assistant", "content": "Hello"}}])
+    )
     litellm.set_verbose = True
     response = litellm.completion(
         model="azure/o1-preview",
         messages=[{"role": "user", "content": "hi"}],
         max_retries=0,
+        api_key="fake-key",
+        api_base="https://fake-azure.openai.azure.com",
+        api_version="2024-10-21",
     )
 
     mock_get_openai_client.assert_called_once()
     assert mock_get_openai_client.call_args.kwargs["max_retries"] == 0
+    assert response.choices[0].message.content == "Hello"
 
 
 @pytest.mark.asyncio
diff --git a/tests/llm_translation/test_azure_openai.py b/tests/llm_translation/test_azure_openai.py
index 0fa72b45ed8..e6528e77749 100644
--- a/tests/llm_translation/test_azure_openai.py
+++ b/tests/llm_translation/test_azure_openai.py
@@ -335,6 +335,10 @@ def test_azure_gpt_4o_with_tool_call_and_response_format(api_version):
     ]
 
     with patch.object(client.chat.completions.with_raw_response, "create") as mock_post:
+        mock_post.return_value.headers = {}
+        mock_post.return_value.parse.return_value = litellm.ModelResponse(
+            choices=[{"message": {"role": "assistant", "content": InvestigationOutput().model_dump_json()}}]
+        )
         response = litellm.completion(
             model="azure/gpt-4.1-mini",
             messages=[
@@ -362,6 +366,7 @@ def test_azure_gpt_4o_with_tool_call_and_response_format(api_version):
             assert "response_format" in mock_post.call_args.kwargs
         else:
             assert "response_format" not in mock_post.call_args.kwargs
+        assert response.choices[0].message.content == InvestigationOutput().model_dump_json()
 
 
 def test_map_openai_params():
diff --git a/tests/llm_translation/test_openai.py b/tests/llm_translation/test_openai.py
index 2b9abdec5d0..af4ba85d58e 100644
--- a/tests/llm_translation/test_openai.py
+++ b/tests/llm_translation/test_openai.py
@@ -292,15 +292,21 @@ class TestOpenAIChatCompletion(BaseLLMChatTest):
 def test_openai_max_retries_0(mock_get_openai_client):
     import litellm
 
+    mock_get_openai_client.return_value.chat.completions.with_raw_response.create.return_value.headers = {}
+    mock_get_openai_client.return_value.chat.completions.with_raw_response.create.return_value.parse.return_value = (
+        ModelResponse(choices=[{"message": {"role": "assistant", "content": "Hello"}}])
+    )
     litellm.set_verbose = True
     response = litellm.completion(
         model="gpt-4o-mini",
         messages=[{"role": "user", "content": "hi"}],
         max_retries=0,
+        api_key="fake-key",
     )
 
     mock_get_openai_client.assert_called_once()
     assert mock_get_openai_client.call_args.kwargs["max_retries"] == 0
+    assert response.choices[0].message.content == "Hello"
 
 
 @patch("litellm.main.openai_chat_completions._get_openai_client")
diff --git a/tests/local_testing/test_completion.py b/tests/local_testing/test_completion.py
index f8f23ea015a..43ed57f63af 100644
--- a/tests/local_testing/test_completion.py
+++ b/tests/local_testing/test_completion.py
@@ -3999,10 +3999,14 @@ def test_completion_novita_ai():
     openai_client = OpenAI(api_key="fake-key")
 
     with patch.object(
-        openai_client.chat.completions, "create", new=MagicMock()
+        openai_client.chat.completions.with_raw_response, "create"
     ) as mock_call:
+        mock_call.return_value.headers = {}
+        mock_call.return_value.parse.return_value = litellm.ModelResponse(
+            choices=[{"message": {"role": "assistant", "content": "Hello"}}]
+        )
         try:
-            completion(
+            response = completion(
                 model="novita/meta-llama/llama-3.3-70b-instruct",
                 messages=messages,
                 client=openai_client,
@@ -4010,6 +4014,7 @@ def test_completion_novita_ai():
             )
 
             mock_call.assert_called_once()
+            assert response.choices[0].message.content == "Hello"
 
             # Verify model is passed correctly
             assert (