Interactive · Grade 10 · Free
Build the context. Then choose the model.
Your goal: get the best speech out of Claude for the least money, by choosing the right context, the right model and the right effort. An AI model knows a lot about the world and nothing about you: not your audience, not which of your sources are trustworthy. Whatever you leave out, it guesses; whatever you put in, it tends to believe. Your job is to get Claude to draft a speech that could persuade school governors to start the day at 10:00, by choosing context from seven buckets (some of it wrong), then choosing which model and how much effort to spend, and watching the cost, the speed and the score.
Opens full screen. No account.
Level 2: your own problem. Once you have played the board, Level 2 lets you type your own prompt about a problem of your own. Your teacher opens it in class and gives you a code. Go to Level 2 →
- context buckets
- 7
- Claude models to compare
- 4
- leaderboards
- 5
Your choices are sent through Tot Academy’s server to Anthropic’s Claude to write the answer. Nothing you choose is stored and there are no accounts. After each run we keep one anonymous line (model, effort, scores, cost, time and an animal nickname) for the leaderboards, with no text and no personal name. The nickname is one of 50 animals worked out from your network, so many people share each one. Privacy notice →
From the people who made these
The tools show you where the marks go. A tutor helps you get them back.
Tot Academy tutors ESAT and university admissions tests, IB, MYP, IGCSE and the Bulgarian curriculum — in Sofia and online. A trial lesson is an hour, and tells us both more than any form.