test(harness): move live OpenAI chat tests to chat_live_openai suite

Per audit 5/8c:
- test_openai.py: live prediction-param pair, TestOpenAIChatCompletion,
  o1 parallel tool calls, web search (+streaming) with the
  validate_web_search_annotations helper, codex (+stream), gemini streaming
  bridge, deepresearch bridge, tool calling, gpt5/gpt-5-codex reasoning,
  n>1 streaming pair, gpt-5 web search; the streaming-handler reasoning
  golden, pdf-url and xhigh-reasoning keepers stay;
  TestOpenAIGPT4OAudioTranscription stays (non-chat, Sameer)
- test_openai_o1.py: TestOpenAIO1, TestOpenAIO3, test_o3_reasoning_effort,
  test_streaming_response moved; the o1 transform goldens
  (system-role/max_completion_tokens mapping, vision support,
  completion_tokens_details) stay
- test_gpt4o_audio.py, test_prompt_caching.py,
  test_router_llm_translation_tests.py moved whole (live, nothing local)

tests/harness_suites/chat_live_openai/test_gpt4o_audio.py

@@ -50,7 +50,6 @@ async def check_streaming_response(completion):
 
 
 @pytest.mark.asyncio
-# @pytest.mark.flaky(retries=3, delay=1)
 @pytest.mark.parametrize("stream", [True, False])
 async def test_audio_output_from_model(stream):
     audio_format = "pcm16"

tests/harness_suites/chat_live_openai/test_openai_o_series_live.py

@@ -0,0 +1,78 @@
+import os
+import sys
+from unittest.mock import patch
+
+sys.path.insert(
+    0, os.path.abspath("../..")
+)  # Adds the parent directory to the system path
+
+
+import pytest
+
+import litellm
+from litellm import ModelResponse
+from base_llm_unit_tests import BaseLLMChatTest
+
+class TestOpenAIO1(BaseLLMChatTest):
+    def get_base_completion_call_args(self):
+        return {
+            "model": "o1",
+        }
+
+    def get_client(self):
+        from openai import OpenAI
+
+        return OpenAI(api_key="fake-api-key")
+
+    def test_prompt_caching(self):
+        """Temporary override. o1 prompt caching is not working."""
+        pass
+
+
+class TestOpenAIO3(BaseLLMChatTest):
+    def get_base_completion_call_args(self):
+        return {
+            "model": "o3-mini",
+        }
+
+    def get_client(self):
+        from openai import OpenAI
+
+        return OpenAI(api_key="fake-api-key")
+
+    def test_prompt_caching(self):
+        """Override, as o3 prompt caching is flaky"""
+        pass
+
+
+def test_o3_reasoning_effort():
+    resp = litellm.completion(
+        model="o3-mini",
+        messages=[{"role": "user", "content": "Hello!"}],
+        reasoning_effort="high",
+    )
+    assert resp.choices[0].message.content is not None
+
+
+@pytest.mark.parametrize("model", ["o1", "o3-mini"])
+def test_streaming_response(model):
+    """Test that streaming response is returned correctly"""
+    from litellm import completion
+
+    response = completion(
+        model=model,
+        messages=[
+            {"role": "system", "content": "Be a good bot!"},
+            {"role": "user", "content": "Hello!"},
+        ],
+        stream=True,
+    )
+
+    assert response is not None
+
+    chunks = []
+    for chunk in response:
+        chunks.append(chunk)
+
+    resp = litellm.stream_chunk_builder(chunks=chunks)
+    print(resp)

tests/llm_translation/test_openai_o1.py

@@ -11,7 +11,6 @@ import pytest
 
 import litellm
 from litellm import ModelResponse
-from base_llm_unit_tests import BaseLLMChatTest
 
 
 @pytest.mark.parametrize("model", ["o1"])
@@ -110,38 +109,6 @@ def test_litellm_responses():
     assert isinstance(response.usage.completion_tokens_details, CompletionTokensDetails)
 
 
-class TestOpenAIO1(BaseLLMChatTest):
-    def get_base_completion_call_args(self):
-        return {
-            "model": "o1",
-        }
-
-    def get_client(self):
-        from openai import OpenAI
-
-        return OpenAI(api_key="fake-api-key")
-
-    def test_prompt_caching(self):
-        """Temporary override. o1 prompt caching is not working."""
-        pass
-
-
-class TestOpenAIO3(BaseLLMChatTest):
-    def get_base_completion_call_args(self):
-        return {
-            "model": "o3-mini",
-        }
-
-    def get_client(self):
-        from openai import OpenAI
-
-        return OpenAI(api_key="fake-api-key")
-
-    def test_prompt_caching(self):
-        """Override, as o3 prompt caching is flaky"""
-        pass
-
-
 def test_o1_supports_vision():
     """Test that o1 supports vision"""
     os.environ["LITELLM_LOCAL_MODEL_COST_MAP"] = "True"
@@ -151,34 +118,3 @@ def test_o1_supports_vision():
             assert v.get("supports_vision") is True, f"{k} does not support vision"
 
 
-def test_o3_reasoning_effort():
-    resp = litellm.completion(
-        model="o3-mini",
-        messages=[{"role": "user", "content": "Hello!"}],
-        reasoning_effort="high",
-    )
-    assert resp.choices[0].message.content is not None
-
-
-@pytest.mark.parametrize("model", ["o1", "o3-mini"])
-def test_streaming_response(model):
-    """Test that streaming response is returned correctly"""
-    from litellm import completion
-
-    response = completion(
-        model=model,
-        messages=[
-            {"role": "system", "content": "Be a good bot!"},
-            {"role": "user", "content": "Hello!"},
-        ],
-        stream=True,
-    )
-
-    assert response is not None
-
-    chunks = []
-    for chunk in response:
-        chunks.append(chunk)
-
-    resp = litellm.stream_chunk_builder(chunks=chunks)
-    print(resp)