CLI erroneously sends unsupported parameters (temperature/top_p) to the o3-mini model

[MrJarnould]

Confirm this is an issue with the Python library and not an underlying OpenAI API

• This is an issue with the Python library

Describe the bug

Using openai api chat.completions.create with the newly released o3-mini-2025-01-31 model triggers errors about unsupported parameters even when those parameters are not explicitly set in the CLI command. Specifically, temperature and top_p appear to be sent to the API, causing 400 errors.

However, if you set --temperature "1" and --top_p "1", then no error is produced and a chat response is obtained.

Note: installed openai using pipx v1.7.1

To Reproduce

1. Use the CLI without specifying --temperature nor --top_p:

❯ openai api chat.completions.create --message "user" "What's the capital of France?" -m "o3-mini-2025-01-31"
Error: Error code: 400 - {'error': {'message': "Unsupported parameter: 'temperature' is not supported with this model.", 'type': 'invalid_request_error', 'param': 'temperature', 'code': 'unsupported_parameter'}}

2. Confirm the same request works using cURL:

❯ curl \
  https://api.openai.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "o3-mini-2025-01-31",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "What'\''s the capital of France?"
          }
        ]
      }
    ],
    "response_format": {
      "type": "text"
    },
    "reasoning_effort": "low"
  }'

{
  "id": "chatcmpl-AwNe73vdAtaqbpdecVpG3XNRwfLbt",
  "object": "chat.completion",
  "created": 1738477283,
  "model": "o3-mini-2025-01-31",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "The capital of France is Paris.",
        "refusal": null
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 17,
    "total_tokens": 29,
    "prompt_tokens_details": {
      "cached_tokens": 0,
      "audio_tokens": 0
    },
    "completion_tokens_details": {
      "reasoning_tokens": 0,
      "audio_tokens": 0,
      "accepted_prediction_tokens": 0,
      "rejected_prediction_tokens": 0
    }
  },
  "service_tier": "default",
  "system_fingerprint": "fp_8bcaa0ca21"
}

3. Observe that the cURL command fails when setting "temperature": 0.5:

❯ curl \
  https://api.openai.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "o3-mini-2025-01-31",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "What'\''s the capital of France?"
          }
        ]
      }
    ],
    "response_format": {
      "type": "text"
    },
    "temperature": 0.5,
    "reasoning_effort": "low"
  }'

{
  "error": {
    "message": "Unsupported parameter: 'temperature' is not supported with this model.",
    "type": "invalid_request_error",
    "param": "temperature",
    "code": "unsupported_parameter"
  }
}

4. Observe that the cURL command works when setting "temperature": 1:

❯ curl \
  https://api.openai.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "o3-mini-2025-01-31",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "What'\''s the capital of France?"
          }
        ]
      }
    ],
    "response_format": {
      "type": "text"
    },
    "temperature": 1,
    "reasoning_effort": "low"
  }'

{
  "id": "chatcmpl-AwNYOeW1akjt6dMLkH3UNSrdXk5tV",
  "object": "chat.completion",
  "created": 1738476928,
  "model": "o3-mini-2025-01-31",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "The capital of France is Paris.",
        "refusal": null
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 17,
    "total_tokens": 29,
    "prompt_tokens_details": {
      "cached_tokens": 0,
      "audio_tokens": 0
    },
    "completion_tokens_details": {
      "reasoning_tokens": 0,
      "audio_tokens": 0,
      "accepted_prediction_tokens": 0,
      "rejected_prediction_tokens": 0
    }
  },
  "service_tier": "default",
  "system_fingerprint": "fp_8bcaa0ca21"
}

5. Observe the CLI still fails when setting --temperature "1", now with a new "Unsupported parameter: 'top_p' is not supported with this model." error:

❯ openai api chat.completions.create --message "user" "What's the capital of France?" -m "o3-mini-2025-01-31" --temperature "1"
Error: Error code: 400 - {'error': {'message': "Unsupported parameter: 'top_p' is not supported with this model.", 'type': 'invalid_request_error', 'param': 'top_p', 'code': 'unsupported_parameter'}}

6. Observe that the CLI request works when setting --temperature "1" --top_p "1":

❯ openai api chat.completions.create --message "user" "What's the capital of France?" -m "o3-mini-2025-01-31" --temperature "1" --top_p "1"
The capital of France is Paris.

Code snippets

OS

macOS

Python version

Python v3.12.5

Library version

openai 1.61.0

[RobertCraigie]

Thanks for the report! This will be fixed in the next release #2078

[csjcoderepo]

Thank you!!

[fateme-hshm96]

Hello. I still get the same error when setting the temperature with o3-mini. Do you have an estimate of when this might get fixed?
Thank you!

[RobertCraigie]

For API level support please go to the community.openai.com forum.

[py-goh]

Hi, I think this same error happened on the python CLI not from the openAI API

If you try to create an assistant with 4-o-mini model: client.beta.assistants.create()

And try to update the model to o3-mini: client.beta.assistants.update()

You will end up getting this error:

{
...
File \"/usr/local/lib/python3.8/dist-packages/openai/resources/beta/assistants.py\", line 355, in update\n    return self._post(\n  File \"/usr/local/lib/python3.8/dist-packages/openai/_base_client.py\", line 1242, in post\n    return cast(ResponseT, self.request(cast_to, opts, stream=stream, stream_cls=stream_cls))\n  File \"/usr/local/lib/python3.8/dist-packages/openai/_base_client.py\", line 919, in request\n    return self._request(\n  File \"/usr/local/lib/python3.8/dist-packages/openai/_base_client.py\", line 1023, in _request\n    raise self._make_status_error_from_response(err.response) from None\nopenai.BadRequestError: Error code: 400 - {'error': {'message': \"Unsupported value: 'reasoning_effort' does not support 'null' with this model.\", 'type': 'invalid_request_error', 'param': None, 'code': 'unsupported_value'}}\n"