The context window and its limits
A model has no memory. Everything it knows about your conversation arrives fresh in every request, and that space has a size, a price and some blind spots.
In this lesson you'll learn to
- Define the context window, and list what has to fit inside it in one request.
- Explain why a model has no memory between calls, and how chat apps fake it by resending history.
- Estimate how input tokens and cost grow over a long conversation, and name ways to keep them down.
- Explain why long context isn't free, what "lost in the middle" means, and how the knowledge cutoff fits in.
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.