I need a feature to tune the thinking parameters due to the trade-off between cost and output tokens.
Please see the relevant post below.