[BUG] OpenAIResponsesModel.structured_output() drops config params (temperature, reasoning, etc.)

[Di-Is]

Checks

• I have updated to the lastest minor and patch version of Strands
• I have checked the documentation and this is not expected behavior
• I have searched ./issues and there are no duplicates of my issue

Strands Version

1.30.0 (main branch HEAD)

Python Version

3.13.7

Operating System

macOS 15.7.4

Installation Method

pip

Steps to Reproduce

from strands.models.openai_responses import OpenAIResponsesModel

model = OpenAIResponsesModel(
    model_id="gpt-5.4",
    params={
        "max_output_tokens": 1000,
        "reasoning": {"effort": "high"},
    },
)

model.structured_output(MySchema, prompt="hello")
# max_output_tokens and reasoning are silently dropped

Expected Behavior

All config params (max_output_tokens, reasoning, instructions, etc.) should be passed through to client.responses.parse(), excluding only model/input (already handled) and stream (injected by _format_request(), incompatible with parse()).

Actual Behavior

Only model, input, and text_format are passed to responses.parse().
All other config params are silently ignored because structured_output() extracts only ["input"] from _format_request() result.

Additional Context

_format_request() correctly spreads self.config["params"] into the request dict (L407), but structured_output() discards everything except input (L363).

This also affects instructions (system_prompt) — it is built by _format_request() but never forwarded to responses.parse().

The stream() method does not have this issue since it passes the full request dict to responses.create().

Possible Solution

Build parse_kwargs from the full _format_request() result, excluding model/input (already handled explicitly) and stream (hardcoded by _format_request(), incompatible with parse()). Other params pass through, letting the API/SDK validate incompatible combinations.