ReasoningParams

class nemo_microservices.types.shared_params.ReasoningParams

Bases: TypedDict

effort: str

Option for OpenAI models to specify low, medium, or high reasoning effort which balances between speed and reasoning accuracy.

end_token: str

Configure the end token to trim reasoning context based on the model’s reasoning API. Used for omitting Nemotron reasoning steps from output denoted with </think> tags