[Bug]: Inconsistent response_format handling between Fireworks AI models

[jtlicardo]

What happened?

When using litellm's completion() function with Fireworks AI models, I discovered inconsistent behavior with the response_format parameter:

What works:

• llama-v3p3-70b-instruct accepts response_format={"type": "text"}
• deepseek-v3 only works with response_format=None

What breaks:

• deepseek-v3 throws an error with response_format={"type": "text"}:
  Error: litellm.BadRequestError: Fireworks_aiException - Error code: 400 - {'error': {'object': 'error', 'type': 'invalid_request_error', 'message': "Extra inputs are not permitted, field: 'response_format.schema_field', value: None"}}

Expected behavior:
The response_format parameter should work consistently across all Fireworks AI models. Either both models should accept {"type": "text"}, or litellm should handle the model-specific differences transparently.

Relevant log output

Give Feedback / Get Help: https://github.com/BerriAI/litellm/issues/new
LiteLLM.Info: If you need to debug this error, use `litellm.set_verbose=True'.


Provider List: https://docs.litellm.ai/docs/providers

Traceback (most recent call last):
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\llms\openai\openai.py", line 657, in completion
    raise e
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\llms\openai\openai.py", line 583, in completion
    self.make_sync_openai_chat_completion_request(
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\llms\openai\openai.py", line 395, in make_sync_openai_chat_completion_request
    raise e
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\llms\openai\openai.py", line 377, in make_sync_openai_chat_completion_request
    raw_response = openai_client.chat.completions.with_raw_response.create(
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\openai\_legacy_response.py", line 356, in wrapped
    return cast(LegacyAPIResponse[R], func(*args, **kwargs))
                                      ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\openai\_utils\_utils.py", line 275, in wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\openai\resources\chat\completions.py", line 829, in create
    return self._post(
           ^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\openai\_base_client.py", line 1280, in post
    return cast(ResponseT, self.request(cast_to, opts, stream=stream, stream_cls=stream_cls))
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\openai\_base_client.py", line 957, in request
    return self._request(
           ^^^^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\openai\_base_client.py", line 1061, in _request
    raise self._make_status_error_from_response(err.response) from None
openai.BadRequestError: Error code: 400 - {'error': {'object': 'error', 'type': 'invalid_request_error', 'message': "Extra inputs are not permitted, field: 'response_format.schema_field', value: None"}}

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\main.py", line 1619, in completion
    raise e
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\main.py", line 1592, in completion
    response = openai_chat_completions.completion(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\llms\openai\openai.py", line 667, in completion
    raise OpenAIError(
litellm.llms.openai.common_utils.OpenAIError: Error code: 400 - {'error': {'object': 'error', 'type': 'invalid_request_error', 'message': "Extra inputs are not permitted, field: 'response_format.schema_field', value: None"}}

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\scratch.py", line 30, in <module>
    response = completion(
               ^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\utils.py", line 994, in wrapper
    raise e
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\utils.py", line 875, in wrapper
    result = original_function(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\main.py", line 2974, in completion
    raise exception_type(
          ^^^^^^^^^^^^^^^
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\litellm_core_utils\exception_mapping_utils.py", line 2190, in exception_type
    raise e
  File "C:\Users\josip\Documents\GitHub\bpmn-assistant\.venv\Lib\site-packages\litellm\litellm_core_utils\exception_mapping_utils.py", line 325, in exception_type
    raise BadRequestError(
litellm.exceptions.BadRequestError: litellm.BadRequestError: Fireworks_aiException - Error code: 400 - {'error': {'object': 'error', 'type': 'invalid_request_error', 'message': "Extra inputs are not permitted, field: 'response_format.schema_field', value: None"}}

Are you a ML Ops Team?

No

What LiteLLM version are you on ?

v1.56.5

Twitter / LinkedIn details

No response