500 Internal Server Error with o3‑mini + file_search on EastUS2 region (Assistants API)

Question

Our application has been successfully creating Azure OpenAI assistants with o3-mini (2025‑01‑31 build), attaching a vector store, enabling file_search, and calling /runs on the Assistants API. As of today, these calls now return a 500 Internal Server Error:

POST https://<your‑endpoint>.openai.azure.com/openai/threads/<thread_id>/runs?api‑version=2024‑05‑01‑preview

500 Internal Server Error

{

"error": {

"message": "The server had an error processing your request. Sorry about that! You can retry your req (truncated...)"

}

}

This flow worked without issue until very recently, with no changes in our code or Azure configuration.

Steps to reproduce:

Create an assistant with model = o3-mini (2025-01-31).

Attach tools including file_search, plus a file-backed vector store.

1. Creates assistant thread, then POSTs to /runs.

The POST returns an HTTP 500 Internal Server Error (no further details).

---

[EDIT]: It looks like error is happening when passing reasoning_effort paramenter when creating assistant. Per OpenAI docs that is possible for o-series models only. I omitted and passed that argument when calling assistant run and seem to work now.

Answer 1

Hi Goran

I am getting partial success in East Us region (not able summarize but not failing at least with 500 internal server errors, managable) and with API version - "2025-04-01-preview"

Attached sample code and results for reference

%pip install openai

from openai import AzureOpenAI
    
client = AzureOpenAI(
    api_key="<yourazureopenaikey>",  
    api_version="2025-04-01-preview",
    azure_endpoint = "https://<yourazureopenairesourcename>.openai.azure.com/"
    )

assistant = client.beta.assistants.create(
  name="Recommender Assistant",
  instructions="You are a bot summarize the document and gives a improved content for attached pdf.",
  model="o3-mini",
  tools=[{"type": "file_search"}],
)



# Create a vector store called "Financial Statements"
vector_store = client.vector_stores.create(name="Financial Statements")
 
# Ready the files for upload to OpenAI
file_paths = ["CSA_Test.pdf"]
file_streams = [open(path, "rb") for path in file_paths]
 
# Use the upload and poll SDK helper to upload the files, add them to the vector store,
# and poll the status of the file batch for completion.
file_batch = client.vector_stores.file_batches.upload_and_poll(
  vector_store_id=vector_store.id, files=file_streams
)
 
# You can print the status and the file counts of the batch to see the result of this operation.
print(file_batch.status)
print(file_batch.file_counts)


# you can replace with vector_store.id instead of string value used here
assistant = client.beta.assistants.update(
  assistant_id=assistant.id,
  tool_resources={"file_search": {"vector_store_ids": [vector_store.id]}},
)


assistant = client.beta.assistants.create(
  instructions="Please summarize the attached document",
  model="o3-mini",
  tools=[{"type": "file_search"}],
  tool_resources={
    "file_search": {
      "vector_store_ids": [vector_store.id]
    }
  }
)


thread = client.beta.threads.create()

run = client.beta.threads.runs.create(
 
  thread_id=thread.id,
  assistant_id=assistant.id
)


#retrieve the results with run
import time

# Looping until the run completes or fails
while run.status in ['queued', 'in_progress', 'cancelling']:
  time.sleep(1)
  run = client.beta.threads.runs.retrieve(
    thread_id=thread.id,
    run_id=run.id
  )

if run.status == 'completed':
  messages = client.beta.threads.messages.list(
    thread_id=thread.id
  )
  print(messages)
elif run.status == 'requires_action':
  # the assistant requires calling some functions
  # and submit the tool outputs back to the run
  print("processing, Please rerun")
  pass
else:
  print(run.status)

Looking forward to your update with Code and result

Note: Have not used reasoning effort here, have modified code from playground instead of file search code, as i faced difficulty in event handler step.

Reference used for API Version

1. https://learn.microsoft.com/en-us/azure/ai-foundry/openai/api-version-lifecycle?tabs=key
2. https://learn.microsoft.com/en-us/azure/ai-foundry/openai/reference-preview

Thank you

Answer 2

Hello Goran !

Thank you for posting on Microsoft Learn.

What you are sharing is a known limitation introduced in the Assistants API (May 2024 preview) when using the o3-mini model with file_search and reasoning_effort in the EastUS2 region.

You're right in identifying that the problem is caused by passing the reasoning_effort parameter when starting a run (/runs endpoint) with a model (o3-mini) that currently does not support it under Azure Assistants API, despite OpenAI general documentation saying it's allowed for o-series models.

reasoning_effort is only supported for chat completions, not for assistant runs using file search with o3-mini.

You can omit the reasoning_effort parameter when creating the assistant or calling /runs for that assistant.

Even though reasoning_effort is accepted by the OpenAI API, Azure Assistants API might still enforce stricter checks or lag slightly behind in compatibility with model features.

Try to use:

{
  "assistant_id": "...",
  "instructions": "Answer using uploaded files",
  "tools": [{"type": "file_search"}]
}