Free IQ test · no sign-up
A free IQ test that tells you what the number means
Ten reasoning questions, an instant result and a written explanation of every answer. No email, no account, no payment step — and no invented precision: a short test gives you a band, so a band is what you get.
The scale you are placed on
Mean 100, standard deviation 15
Half of everyone lands between 90 and 110. The tails look dramatic and are very thin.
How it works
Three steps, and none of them involve your email address
Most free IQ tests are free up to the moment the score appears. This one is free at the moment the score appears, which is the only moment that counts.
Answer ten items
Five kinds of reasoning, no timer and no penalty for going back. Each figure is drawn as vector graphics, so it stays legible on a phone.
See a band, not a fake number
A ten-item sample supports a band and a percentile range. You get both, together with how you did on each type of question.
Read why each answer is the answer
Every item comes with the rule that generates it. Items that only work through ambiguity do not survive that step.
What is on the test
Five kinds of reasoning, and one deliberate omission
Reasoning tests differ mostly in what they refuse to ask about. Everything below was chosen because performance on it moves with general reasoning ability rather than with how much you happen to have read.
Abstract reasoning
Matrix items where the rule lives in the relationship between rows and columns. This is the part of a test that comes closest to measuring reasoning free of prior knowledge.
Numerical reasoning
Number series where the pattern sits in the differences, the ratios or a simple generating formula. Arithmetic is deliberately easy so that the difficulty is in spotting the rule.
Verbal reasoning
Analogies that test whether you can name the relationship between two words and carry it across to a second pair, rather than whether you happen to know an unusual word.
Spatial reasoning
Rotation and reflection tasks. Telling a turn from a flip is one of the cleanest single indicators in the whole reasoning literature and one of the easiest to get wrong under pressure.
Logical reasoning
Short formal arguments where the job is to say what follows and, more importantly, what does not. Most wrong answers here are statements that are plausible rather than entailed.
What is deliberately left out
No general knowledge, no vocabulary obscurity, no mental arithmetic under a clock and no cultural trivia. Those things inflate a score without telling you much about reasoning.
The basics
What an IQ score actually is
An IQ score is not a measurement of a quantity the way your height or your resting heart rate is. Nobody has ever measured a unit of intelligence. What an IQ test measures is your position relative to other people who took the same test, converted onto a scale that was chosen for convenience.
That scale has a mean of 100 and a standard deviation of 15. Both numbers are conventions. The mean could have been 500 and the standard deviation 100, which is roughly what the SAT does with the same underlying idea. Setting the mean at 100 means that a score of 100 is defined as the middle of the comparison group, not as a particular amount of reasoning power.
Two consequences follow, and almost every misunderstanding about IQ comes from missing one of them.
The first is that an IQ score is meaningless without knowing who you were compared against. A score of 115 against a representative national sample and a score of 115 against a self-selected group of people who clicked a link on a puzzle forum are not the same thing. When a site will not tell you what its norm group is, the number it hands you is decoration.
The second is that the scale is not linear in any everyday sense. The gap between 100 and 115 covers about 34 percent of the population. The gap between 145 and 160, which is the same fifteen points, covers about one person in every two thousand. Points get rarer as you move outwards, which is why differences at the top of the scale are both impressive-sounding and very hard to measure reliably.
An IQ score is a rank dressed up as a number, and the rank is only as good as the group it was taken against.
Reading a score without over-reading it
Because the scale is built on the normal distribution, a score translates directly into a percentile — the share of the comparison group you scored at or above. That translation is fixed and does not depend on the test, as long as the test really uses a mean of 100 and a standard deviation of 15.
| IQ score | Percentile | Rarity | Usual band |
|---|---|---|---|
| 70 | 2.3% | about 1 in 44 | Very low |
| 80 | 9.1% | about 1 in 11 | Low average |
| 85 | 15.9% | about 1 in 6 | Low average |
| 90 | 25.2% | about 1 in 4 | Average |
| 100 | 50.0% | about 1 in 2 | Average |
| 110 | 74.8% | about 1 in 4 | High average |
| 115 | 84.1% | about 1 in 6 | High average |
| 120 | 90.9% | about 1 in 11 | Very high |
| 130 | 97.7% | about 1 in 44 | Extremely high |
| 140 | 99.62% | about 1 in 261 | Extremely high |
The rarity column is where intuition usually fails. A score of 130 sounds like a comfortable distance above 120, and in percentile terms it is: one person in fifty instead of one in eleven. Moving another ten points to 140 takes you to roughly one in two hundred and sixty. The scale stretches as you climb it.
The full table, together with a calculator that converts any score, is on the IQ percentile chart.
What the bands are called
Test publishers attach descriptive labels to score ranges. The labels have changed repeatedly over the last century, and the direction of travel has been steadily away from words that sound like verdicts about a person.
| Score range | Current label | Share of people | Percentile range |
|---|---|---|---|
| 130 and above | Extremely high | 2.2% | 98th percentile and up |
| 120–129 | Very high | 6.7% | 91st–97th percentile |
| 110–119 | High average | 16.1% | 75th–90th percentile |
| 90–109 | Average | 50% | 25th–74th percentile |
| 80–89 | Low average | 16.1% | 9th–24th percentile |
| 70–79 | Very low | 6.7% | 2nd–8th percentile |
| 69 and below | Extremely low | 2.2% | Below the 2nd percentile |
These bands come from the Wechsler tradition, which is the one most English-language tests follow. The percentages are not survey results; they are what the normal curve produces by construction. If a test reports a distribution noticeably different from this, either the test is not normed on a representative sample or it is not really using a fifteen-point standard deviation.
We trace how the labels changed — and why moron, defective and genius were all once technical terms — on IQ classification ranges.
The part most sites leave out: the error band
Every test score is a single sample of a noisy process. Psychometricians quantify that with the standard error of measurement, and on a good clinical instrument it is around three points, which produces a 95 percent confidence interval of roughly plus or minus five to six points. On a short unsupervised online test it is considerably wider.
In practice that means a result of 118 and a result of 112 are the same result. Treating a six-point difference between two people, or between two sittings by the same person, as a real difference is the single most common error in how these numbers get used. Our ten-item test reports a band rather than a point precisely because a point would be a false claim about its own precision.
A ten-question test can reliably tell you which quarter of the distribution you are in. It cannot tell you whether you are a 117 or a 123. Any site that prints one of those at you after ten questions has chosen a nicer-looking result over a true one.
Reference library
The tables people actually come here for
Every chart below is generated from the normal distribution rather than copied from another site, and each one states its assumptions on the page.
IQ percentile chart
Every score from 40 to 160 converted to a percentile and a rarity figure, plus a calculator that does any value in between.
Read more →IQ classification ranges
What each band is called today, what it was called in 1916, and why the wording keeps changing.
Read more →What is the average IQ?
Why the answer is 100 by construction, what the Flynn effect did to that, and what an average genuinely tells you.
Read more →What is a good IQ score?
The honest answer depends on the question behind it. Six different versions of “good”, each with its own threshold.
Read more →Average IQ by age
How age norms work, why a raw score peaks in your twenties while your IQ does not, and what changes after sixty.
Read more →Average IQ by country
The dataset everyone quotes, the sampling problems nobody mentions, and how much weight the figures can carry.
Read more →Test types
Mensa, Raven, WAIS, Wonderlic — which is which
People use “IQ test” for half a dozen unrelated instruments with different purposes, formats and levels of evidence behind them.
The Mensa IQ test
What the entry threshold really is, which prior results are accepted, and why no online test qualifies you.
Read more →Raven's Progressive Matrices
The best-known non-verbal reasoning test, the rule families its items are built from, and where the culture-fair claim holds.
Read more →WAIS — the adult clinical standard
Fifteen minutes on what a psychologist actually does across a two-hour session, and what the index scores mean.
Read more →Straight answer
How accurate is a free online IQ test?
Accurate enough to be interesting, nowhere near accurate enough to be official, and the gap between those two statements is where most of the marketing in this niche lives.
Four things separate a web test from a clinical one, and none of them are fixable by writing better questions.
Nobody is watching. A psychologist administering a WAIS controls the room, the timing, the instructions and the number of attempts. A web page controls none of that. You might be tired, interrupted, on your third attempt, or helping a friend with theirs.
The comparison group is self-selected. Clinical tests are normed on stratified samples matched to census data. A web test is normed, at best, on people who searched for an IQ test and finished it — a group that is younger, more online and more curious about its own results than the population is.
Short tests are noisy. Reliability rises with length. Ten items is enough to place you in a broad band; thirty narrows it usefully; the sixty to ninety minutes of a full clinical battery is what it takes to get the error down to a few points.
The format is not the ability. Practice with matrices improves your matrix score faster than it improves your reasoning. On a first attempt this matters little. By your fourth online test it matters a great deal.
None of that makes the exercise pointless. A free test is a sensible first read, a decent way to find out whether the format suits you, and a reasonable filter for deciding whether a supervised test is worth booking. It is simply not evidence, and a site that implies otherwise is selling something.
Reviews
Online IQ tests, compared side by side
What each one costs, what it withholds, how long it runs and what it tells you about its own norms. We say where each one is genuinely the better choice.
IQScore vs iq-test.io
A $6.99 one-off unlock against a test that charges nothing. Where the paid pack is worth it and where it is not.
Read more →MyIQTested vs iq-test.io
The strongest content library in the niche, wrapped around a test that sells the certificate separately.
Read more →123test vs iq-test.io
Twenty years of history and eleven languages, but the free version is a demo and the real test is paid.
Read more →From the blog
Research without the hype
Evidence
How to increase your IQ: what the evidence supports and what it does not
Brain-training apps, dual n-back, supplements, music, sleep. Sorting the interventions with evidence behind them from the ones with marketing behind them.
11 min read
Methods
How accurate is an IQ test, really?
Reliability, validity and the standard error of measurement, explained without the jargon and with the numbers that actually apply.
10 min read
Research
The Flynn effect: why IQ scores rose for a century, and why they may be falling
Raw scores climbed roughly three points a decade for most of the twentieth century. Nobody fully agrees why, and in several countries the climb has stopped.
9 min readQuestions
Free IQ test: frequently asked questions
Twenty of the questions we get most often, answered without the marketing layer.
Is this IQ test really free?
Yes. The ten reasoning questions, the result band, the breakdown by question type and the worked explanation of every item are all free, with no account, no email address and no card details at any point. There is nothing held back behind a payment step, because there is no payment step.
How long does the test take?
Most people finish in six to nine minutes. There is no countdown, and nothing is scored on speed, so working carefully costs you nothing.
Will I get an exact IQ number at the end?
No, and you should be suspicious of any ten-question test that gives you one. Ten items cannot pin a score down to a single point — the measurement error on a sample that short is wider than most of the bands people care about. You get a band and a percentile range instead, which is the honest resolution of a test this length. For a longer sitting with a tighter band, take the 33-question test at iq-test.io.
Do I need to register or give an email address?
No. Nothing is stored on a server and nothing is emailed to you. Your answers stay in your browser for as long as the page is open and disappear when you close it.
What does the test actually measure?
Five kinds of reasoning: abstract pattern completion, number series, verbal analogy, spatial rotation and formal logic. Between them they load fairly heavily on what psychometricians call fluid reasoning — the ability to work out a rule you have never been taught. They do not measure knowledge, vocabulary depth, memory span or processing speed.
Is an online IQ test accurate?
It is accurate enough to place you in a broad band and far too loose for anything official. Nobody checks who is sitting at the keyboard, nobody controls for noise, tiredness or a second attempt, and the comparison group is whoever happened to visit the site rather than a representative sample of the population. We go through the limits in detail on are online IQ tests accurate.
What is considered a good IQ score?
Anything from 90 to 109 is squarely average and describes half the population. From 110 up you are in the top quarter, and 130 — the usual cut for high-IQ societies — is reached by roughly one person in fifty. We unpack why “good” depends entirely on the question you are asking on what is a good IQ score.
Why is the average IQ always 100?
Because it is set there. IQ is not a count of anything; it is a position on a curve. Test publishers take a representative sample, look at the raw score distribution, and define the middle of that distribution as 100 with a standard deviation of 15. If the population got better at the test tomorrow, the test would be restandardised and the average would be 100 again.
Can I retake the test?
Yes, as many times as you like. Be aware that a second attempt on the same ten items is no longer a test of reasoning — it is a test of memory. Practice effects on short reasoning tests are real and can be worth several points.
Is the test suitable for children?
Does the test work on a phone?
Yes. Every figure is drawn as vector graphics that stay sharp at any size, the answer buttons are sized for thumbs, and nothing requires hovering or a mouse. Most people take it on a phone.
Will a high score get me into Mensa?
No online test will, including this one and every other one you will find. Mensa admits members through its own supervised test or through prior results from a short list of approved instruments. A free web test is a reasonable way to decide whether it is worth booking one — nothing more. See the Mensa IQ test guide.
Do you explain the answers?
Every item on the test has a written explanation of the rule behind it, shown after you finish. If an item only works because of a trick, the explanation would show it, which is one of the reasons we write them.
Why are there only ten questions?
Because ten is honest about what it can do and short enough that people actually finish. A longer test genuinely narrows the band — that is the trade-off, and it is why we point people to a longer sitting once they know they are interested.
Do you use the results for anything?
No. There is no analytics event carrying your answers, no server call when you finish, and no database behind this page. The scoring happens in your browser.
Is IQ fixed for life?
Rank order is fairly stable from late childhood onwards, but individual scores move, sometimes by a lot. Schooling, health, sleep, practice with the test format and simple measurement noise all shift the number. Treat any single result as one observation rather than a fixed attribute.
What is the difference between IQ and intelligence?
IQ is a score on a particular test, standardised against a particular group, at a particular time. Intelligence is the thing the test is trying to sample. The two are correlated well enough for the score to be useful and badly enough that it should never be the last word about anyone.
Can I improve my IQ score?
You can improve your score on a specific test format quite easily with practice, which is not the same as improving the underlying ability. The evidence for lasting gains from commercial brain-training is weak; the evidence that additional schooling raises measured IQ is much stronger. We go through what holds up on how to increase IQ.
Which IQ test is the most accurate one available?
For anything that matters, an individually administered clinical test given by a psychologist — the WAIS for adults, the WISC for children. Among unsupervised online tests, accuracy varies mostly with length and with whether the site is honest about its norms. Our side-by-side reviews compare what each one gives you.
Are the questions the same every time?
Yes. The bank is fixed and deterministic, which means every visitor sees the same ten items in the same order. That makes results comparable between people and lets anyone check our working on a specific item.
Ready when you are
Ten questions, no timer, no sign-up, and an explanation for every answer at the end.