An IQ test for geniuses, and an honest word about the top end

There is no official genius score. The label comes from Lewis Terman, who in 1916 described scores above 140 as "near genius or genius". Modern tests do not use the word at all. Wechsler calls 130 and above "very superior" or "extremely high", and stops there.

What there is, is rarity. On the standard scale, with an average of 100 and a standard deviation of 15, each step up the top of the curve is much rarer than the last.

So "genius level" in everyday use usually means about 140, the 99.6th percentile, or 145, about 1 in 741. Both are rare. Neither is a title a test can award. In the history of science, the word genius is usually earned by a body of work, not a number, which is why the famous genius IQs you see online are estimates, such as Einstein's.

These three items are harder than most of the test: one stacks four rules, one uses a bigger grid, one adds a total to the pattern. They are written for this page and are not test questions. If you solve all three in under two minutes, the hardest items in the test should suit you.

Measuring very high IQ is hard for three plain reasons.

• Too few people to compare with. A test is scored against a norm sample. If a sample has 2,000 people, only about five would be expected at 145 or above. Scores up there lean on the shape of the curve, not on people actually tested.
• Too few hard items. To tell 145 from 155 a test needs many items that only those people can solve. Most tests have a handful, so one lucky or unlucky answer moves the score a lot.
• The ceiling. Every test has a highest possible score. The WAIS-IV reports Full Scale IQs up to 160. A perfect score on a short online test does not mean "at least 160", it means "above what this test can see".

Self-described "high range" tests try to solve this with very hard untimed puzzles. Most are self-administered with no supervision, and none has the kind of large, representative norm sample that clinical tests use. High-IQ societies are cautious about them for that reason. A short timed test like ours is honest about its limit instead: it separates well up to about 130 and gives a rougher read above that.

If you want a high score that counts for something, these societies set the bar. Each accepts only supervised tests or approved prior scores, not online tests like this one.

The Mensa practice test guide explains which tests Mensa accepts. If you have old SAT, GRE or LSAT scores, some societies accept those too; the score converters show what they imply.

Two long studies are worth knowing about.

Terman's Genetic Studies of Genius began in 1921 and followed 1,528 Californian children with very high scores for the rest of their lives. Most did well: good health, good jobs, stable lives. Very few became famous for original work. A high score predicted a good life far better than it predicted genius.

The Study of Mathematically Precocious Youth has followed thousands of people found in the top 1% of reasoning at age 12 or 13 since 1971. Even inside that top 1%, people with higher scores were more likely, decades later, to earn doctorates, patents and published research. Ability differences at the top do matter, in other words. They are just not the whole story: interests, effort and opportunity shape what people do with them.

If you want the trait side of the question, the Am I a genius? and Am I gifted? self-checks cover it. They describe how you think. A test measures how you reason.