Support for gpt5's 'verbosity' option

[ryx2]

Description

from openai import OpenAI
client = OpenAI()

response = client.responses.create(
  model="gpt-5",
  input=[
    {
      "role": "user",
      "content": [
        {
          "type": "input_text",
          "text": "..."
          "type": "output_text",
          "text": "..."
        }
      ]
    }
  ],
  text={
    "format": {
      "type": "text"
    },
    "verbosity": "low"    < --- here
  },
  reasoning={
    "effort": "high",
    "summary": "detailed"
  },
  tools=[],
  store=True
)

References

No response

[gusfraser]

Also think that it should fail gracefully e.g. if temperature is provided - see my related issue raised:

#2482

[Kludex]

It seems we should support openai_text_verbosity or openai_verbosity on the OpenAIResponsesModelSettings.

@DouweM Do we set the text format somewhere, and it would actually fit somewhere else?

[DouweM]

@Kludex The text field on the responses API is also used for the JSON output format/schema:

pydantic-ai/pydantic_ai_slim/pydantic_ai/models/openai.py

Lines 796 to 804 in 5745402

	text: responses.ResponseTextConfigParam | None = None
	if model_request_parameters.output_mode == 'native':
	output_object = model_request_parameters.output_object
	assert output_object is not None
	text = {'format': self._map_json_schema(output_object)}
	elif (
	model_request_parameters.output_mode == 'prompted' and self.profile.supports_json_object_output
	): # pragma: no branch
	text = {'format': {'type': 'json_object'}}

I think the ModelSettings are a good place to put it, but we'd only send it if we're not in one of those JSON branches.

[ryx2]

ok @DouweM i gave claude your instructions and it made #2493 💯