OpenAI Reasoning Levels

OpenAI models now have couple of reasoning levels:
https://developers.openai.com/api/docs/models

It seems that currently only `low`, `medium` and `high` reasoning levels are supported, can we support other reasoning levels (especially `none`)?

Start with the OpenAI model documentation linked in the issue, then locate the Go integration that currently handles the low, medium, and high reasoning levels. Trace the validation or request-building path and its tests, if present. Done means the supported reasoning levels include the requested additional values, especially none, with coverage for the new behavior.

Written by the indexing model from the issue text.