diff --git a/backend/conftest.py b/backend/conftest.py
index 64b912e1..d9ba2607 100644
--- a/backend/conftest.py
+++ b/backend/conftest.py
@@ -1,28 +1,9 @@
 # conftest.py
 import pytest
-import logging
 import os
-from unittest.mock import patch
-
-from openai import AsyncOpenAI
-
-logger = logging.getLogger(__name__)
 
 
 @pytest.hookimpl(tryfirst=True)
 def pytest_configure(config):
     # Set an environment variable to indicate pytest is running
     os.environ["PYTEST_RUNNING"] = "1"
-    os.environ["OPENAI_API_KEY"] = "fake_key"
-
-
-class MockOpenAI(AsyncOpenAI):
-    pass
-
-
-@pytest.fixture(autouse=True)
-@patch("src.llm.OpenAI.client")  # Ensure this matches the import path in your module
-def mock_async_openai():
-    with patch("openai.AsyncOpenAI", MockOpenAI) as mock_async_openai:
-        yield mock_async_openai
-    print("AsyncOpenAI patched")
diff --git a/backend/src/llm/openai.py b/backend/src/llm/openai.py
index 461ea533..435eaa20 100644
--- a/backend/src/llm/openai.py
+++ b/backend/src/llm/openai.py
@@ -18,7 +18,6 @@ def remove_citations(message: Text):
 
 
 class OpenAI(LLM):
-    client = AsyncOpenAI(api_key=config.openai_key)
 
     async def chat(self, model, system_prompt: str, user_prompt: str, return_json=False) -> str:
         logger.debug(
@@ -27,7 +26,8 @@ async def chat(self, model, system_prompt: str, user_prompt: str, return_json=Fa
             )
         )
         try:
-            response = await self.client.chat.completions.create(
+            client = AsyncOpenAI(api_key=config.openai_key)
+            response = await client.chat.completions.create(
                 model=model,
                 messages=[
                     {"role": "system", "content": system_prompt},
@@ -57,16 +57,17 @@ async def chat_with_file(
         files_by_path: Optional[list[LLMFileFromPath]] = None,
         files_by_stream: Optional[list[LLMFileFromBytes]] = None
     ) -> str:
+        client = AsyncOpenAI(api_key=config.openai_key)
         file_ids = await self.__upload_files(files_by_path, files_by_stream)
 
-        file_assistant = await self.client.beta.assistants.create(
+        file_assistant = await client.beta.assistants.create(
             name="ESG Analyst",
             instructions=system_prompt,
             model=model,
             tools=[{"type": "file_search"}],
         )
 
-        thread = await self.client.beta.threads.create(
+        thread = await client.beta.threads.create(
             messages=[
                 {
                     "role": "user",
@@ -79,11 +80,11 @@ async def chat_with_file(
             ]
         )
 
-        run = await self.client.beta.threads.runs.create_and_poll(
+        run = await client.beta.threads.runs.create_and_poll(
             thread_id=thread.id, assistant_id=file_assistant.id
         )
 
-        messages = await self.client.beta.threads.messages.list(thread_id=thread.id, run_id=run.id)
+        messages = await client.beta.threads.messages.list(thread_id=thread.id, run_id=run.id)
 
         message = messages.data[0].content[0].text
 
@@ -95,6 +96,7 @@ async def __upload_files(
         files_by_path: Optional[list[LLMFileFromPath]],
         files_by_stream: Optional[list[LLMFileFromBytes]]
     ) -> list[str]:
+        client = AsyncOpenAI(api_key=config.openai_key)
         if not files_by_path:
             files_by_path = []
         if not files_by_stream:
@@ -103,7 +105,7 @@ async def __upload_files(
         file_ids = []
         for file in files_by_stream + files_by_path:
             logger.info(f"Uploading file '{file.file_name}' to OpenAI")
-            file = await self.client.files.create(
+            file = await client.files.create(
                 file=file.file_path if isinstance(file, LLMFileFromPath) else file.file_stream,
                 purpose="assistants"
             )
diff --git a/backend/tests/llm/openai_test.py b/backend/tests/llm/openai_test.py
index cc10276a..1d230923 100644
--- a/backend/tests/llm/openai_test.py
+++ b/backend/tests/llm/openai_test.py
@@ -49,16 +49,18 @@ class MockListResponse:
 
 
 @pytest.mark.asyncio
-@patch("src.llm.OpenAI.client")  # Ensure this matches the import path in your module
-async def test_chat_with_file_removes_citations(mock_client):
-    mock_client.files.create = AsyncMock(return_value=MockResponse(id="file-id"))
-    mock_client.beta.assistants.create = AsyncMock(return_value=MockResponse(id="assistant-id"))
-    mock_client.beta.threads.create = AsyncMock(return_value=MockResponse(id="thread-id"))
-    mock_client.beta.threads.runs.create_and_poll = AsyncMock(return_value=MockResponse(id="run-id"))
-    mock_client.beta.threads.messages.list = AsyncMock(return_value=MockListResponse)
+@patch("openai.AsyncOpenAI")  # Ensure this matches the import path in your module
+async def test_chat_with_file_removes_citations(mock_async_openai):
+    mock_instance = mock_async_openai.return_value
 
-    llm = OpenAI()
-    response = await llm.chat_with_file(
+    mock_instance.files.create = AsyncMock(return_value=MockResponse(id="file-id"))
+    mock_instance.beta.assistants.create = AsyncMock(return_value=MockResponse(id="assistant-id"))
+    mock_instance.beta.threads.create = AsyncMock(return_value=MockResponse(id="thread-id"))
+    mock_instance.beta.threads.runs.create_and_poll = AsyncMock(return_value=MockResponse(id="run-id"))
+    mock_instance.beta.threads.messages.list = AsyncMock(return_value=MockListResponse)
+
+    client  = OpenAI()
+    response = await client .chat_with_file(
         model="",
         user_prompt="",
         system_prompt="",