Which AI are you?

Want to know which AI is actually best for each task? Read our Claude vs ChatGPT vs Gemini comparison or discover curiosities you didn't know.

What this quiz measures (and what it doesn't)

Let's start with the honest part: this quiz is a game. There is no evidence that the way you organise your downloads folder predicts which language model suits you, and any test promising otherwise is selling you something. What it does do is something more modest and considerably more fun: use eight questions about how you work as an excuse to talk about the real differences between the AI systems on the market today.

Because those differences do exist, and they're more about character than people assume. Current models aren't distinguished mainly by how clever they are — on ordinary day-to-day tasks the big three are reasonably level — but by how they behave when the situation isn't clean. Some prefer to admit they don't know; others take a swing at an answer. Some ask for clarification before starting; others dive in and correct later. Some write with more nuance and less enthusiasm; others the reverse. None of these stances is better in the abstract: it depends what you're using it for.

How you actually choose a model

If you came this far looking for a serious answer, here it is: the real choice comes down to four questions, and none of them is about your personality.

How much does your task care about being exactly right? If you work with data, contracts or code that runs, you need a model that errs conservative and tells you when it isn't sure. If you're generating campaign ideas, that same caution gets in the way and you'd rather have one that takes risks.

Is the task one shot or many steps? Writing an email is one shot, and almost any model handles it. Refactoring a module, reading twenty documents and cross-referencing them, or running a process with tools is long work — and that's where the gaps between models widen sharply. Not losing the thread at step fifteen is where they genuinely separate.

Where does your work live? The best model in the world, in a tab you never open, loses to a middling one embedded where you already are. Living in Gmail and Docs, living in a terminal, and living in a phone app are three different answers.

How much will you use it? Occasional use is covered by any free tier. Heavy daily use makes the price gap between tiers noticeable. And automated use changes the question entirely, because now you're talking about APIs and tokens.

Those four answers translate into a concrete recommendation far better than any personality test. The model selector turns them into a direct result, the Claude vs ChatGPT vs Gemini comparison works through them criterion by criterion, and the 2026 AI models table has the numbers.

Why these quizzes work so well

A note for the curious: personality quizzes hook people through a phenomenon documented since the 1940s, the Barnum effect. Descriptions that are general enough — "you're analytical but you also have an impulsive side" — are accepted as strikingly accurate by almost everyone, because almost everyone recognises themselves in them. Add an identity component ("I'm the kind of person who…") and the result shares itself.

Knowing this doesn't spoil the fun. It's just worth keeping in mind before making work decisions off the back of eight questions. If you want more from this family of content, AI curiosities has facts considerably less intuitive than this one, and the glossary puts the terms that appear everywhere, unexplained, into plain language.