L

Learn AI

Track progress · learn offline

Open

Lesson 2 of 9

Picking the right model

6 min read

You open the assistant and there's a little dropdown: a "fast" model, a "thinking" one, maybe a few with cryptic names. Most people pick one at random and never touch it again. But that dropdown is the single biggest lever you have — and choosing well takes about two seconds.

Not one AI, but a family

Behind that menu isn't one AI wearing different hats — it's a family of separate models, built to trade off against each other. A bigger model is smarter but slower and costs more to run. A smaller one is quick and cheap but flubs the hard stuff. There's no single "best" — only best for this task. So the menu isn't clutter; it's you choosing where to spend.

The model menu is a set of genuinely different models, each a trade-off between how capable, how fast, and how expensive it is. There's no free lunch — more capability costs speed and money.

The one dial worth knowing

You can ignore most of the menu and watch a single distinction: fast versus thinking. A fast model answers the instant you hit enter — perfect for quick questions, rewrites, everyday chat. A thinking model pauses first and works the problem through step by step before replying — slower, but far better at anything with several moving parts: a tricky logic puzzle, a plan with constraints, careful maths. Reach for it when the fast one gets it wrong.

For everyday questions, use the fast model. When a task has several steps or a right answer that's easy to get wrong, switch to a thinking model and accept the wait.

A simple habit: start with the fast model. If its answer is thin or wrong on something that clearly needs care, ask the same thing again with the thinking model. You'll feel the difference immediately.

The gist

You need to draft a quick reply to a friendly email. Which model, and why?

Continue in the app

Take the whole Working with AI course — tracked

Get your personalized path, progress and streaks in the app — this lesson and every next one, in order.