Translates a reasoning level to the OpenAI request fragment to merge into settings. Returns {} when reasoning is not requested, unsupported, or disabled — so the caller can spread it unconditionally.
settings
{}
true for the Responses API (reasoning: { effort }, supports minimal); false for Chat Completions (reasoning_effort, low|medium|high only).
true
reasoning: { effort }
minimal
false
reasoning_effort
low
medium
high
false for models without reasoning support — returns {}.
Translates a reasoning level to the OpenAI request fragment to merge into
settings. Returns{}when reasoning is not requested, unsupported, or disabled — so the caller can spread it unconditionally.