What is my IQ? Find out in 12 minutes

There is no IQ written down anywhere waiting to be looked up. You get one by being tested, or you estimate one from a test you have already taken. The routes differ in time, cost and how far you can trust the number, so here they are side by side.

If you want a number for a formal purpose, such as joining Mensa, a school placement or a clinical question, go straight to a supervised test. If you want to know roughly where you stand and see how you reason, start with the test above and read the result as a range. Our guide to how to find out your IQ compares the five routes in more depth.

These are written for this page and are not test questions. The first is easy; the second is the kind that separates scores at the top. Choose an answer to see the explanation.

Every modern IQ scale puts the average at 100 and one standard deviation at 15 points. That fixes what each score means, whichever test it came from. Here is where common scores sit among 100 people.

Two thirds of people score between 85 and 115, so a result anywhere in that band is ordinary in the best sense. For any score, the IQ percentile calculator gives the exact percentile and the IQ chart shows every level on one page. If you want to know how the number is produced, see how IQ is calculated.

Admissions tests are not IQ tests, but they correlate strongly with general reasoning ability: the SAT correlated .82 with a measure of general ability in Frey and Detterman's 2004 study. That is enough for a rough range, not a precise number. These ranges come from our published conversion models.

Notice how wide the ranges are, and that very high test scores point to IQs a little closer to 100 than you might expect. That is regression to the mean, and it is honest. The SAT to IQ and ACT to IQ converters work for any score and explain the assumptions, and all the converters cover the GRE, LSAT, GMAT and more.

Search "what is my IQ" and take three tests, and you may get three numbers 20 points apart. The reasons are mostly about scoring, not about you.

• Who you are compared with. A professional test compares you with a census-matched norm sample. Many online tests compare you with their own visitors, or with nobody at all.
• Flattering scales. A test that tells most people they are above 120 is more pleasant than accurate. If nearly everyone you know scored high on the same site, that is the reason.
• Length. Fewer questions mean a wider margin of error. A 10-question quiz cannot separate 115 from 130.
• What is tested. Visual puzzles, vocabulary and mental arithmetic measure overlapping but different abilities, so a strong talker and a strong puzzler can swap places between tests.

Our test is 25 visual reasoning puzzles scored on the standard scale. It is not a clinical assessment, and we say what it does not measure. The guide to online IQ test accuracy shows how to judge any test in two minutes, including ours.

Even the best clinical tests report a confidence interval. On the WAIS-5, whose full-scale reliability is about .97, the 95% interval is roughly 5 points either side of the score. A shorter test is less reliable, so its band is wider. Two people who score 118 and 124 on a short test have not been shown to differ.

If you take a second test soon after the first, expect a slightly higher score from familiarity alone. To see real change, leave months between attempts. The IQ confidence interval calculator shows the band around any score, and why IQ scores differ explains the gaps between tests.