Reasoning controls
The same model can answer in one second for a fraction of a cent, or think for a minute and cost fifty times more. You choose which, on every request, with one setting most people leave on the default.
In this lesson you'll learn to
- Describe effort levels and thinking budgets, and set them in a request.
- Predict what more reasoning does to quality, latency and cost.
- Handle thinking in responses, max tokens and streaming correctly.
- Pick an effort level per task and confirm it with a small measurement.
Create a free account to open this lesson
It's free. You'll get every free lesson, quizzes that save your progress, daily flashcards and streaks.