Search for a free IQ test and most of what comes back wants your email before it shows you anything. This one works the other way around. Answer 33 reasoning problems and your score, percentile and four ability readings appear the second you finish. The questions come from ICAR, a public item set built and tested by university researchers, not written to fill a page.
No sign-up, no payment, no result held back at the end. Scoring runs inside your browser, so your answers stay on your device and reach nobody but you.
The test opens on a clean page of its own. You work through four blocks of problems at whatever pace suits you, with no clock and nobody watching over your shoulder.
One button and you are in. Thirty-three problems, a quiet screen, and your full result waiting the moment you place your last answer.
Open the testfour domains · scored on your device · nothing sent to a server
Plenty of sites call themselves a free IQ test. Here is what actually sits behind this one.
Every item comes from the International Cognitive Ability Resource, a public bank that academics have tested on tens of thousands of people and checked against full clinical batteries. Each one carries a research history, which is more than most quiz-style tests can claim.
A two-question test tells you nothing; a two-hour one loses you halfway. Thirty-three items across four domains is the middle ground, long enough to trust the number and brief enough to finish in one sitting.
You land on question one, not on a registration form. No email, no account, and no paywall waiting to spring the moment you reach for your result.
Since the grading happens on your own device, there is nothing to upload and nothing to load. Your last click brings up the IQ figure, the percentile, and all four domain scores together.
The 33 questions fall into four groups, and each leans on a different part of the broad ability psychologists label g.
Grids of shapes with one cell left blank, and your task is to work out what belongs there. These matrix puzzles come closest to a direct reading of fluid intelligence, so the test leans on them hardest: twelve items, more than any other block.
Word meanings, and the relationships that tie them together. Where the matrix puzzles reward thinking on the spot, this block rewards what you have already absorbed over the years: your vocabulary and your feel for how ideas connect. Seven items.
A run of numbers or letters that follows some rule, with the next term left for you to supply. Finding it calls on quick pattern-spotting and a little mental arithmetic held together in working memory. Seven items.
One shape is shown, then four more, and a single one of them is that same object turned to a new position. Spotting it means rotating the figure in your head. This is the most visual block, and the final seven items.
The uneven split is on purpose. Matrix puzzles carry the most weight because they predict general ability better than anything else on the test; the other three blocks stop the score from hinging on one narrow talent and turn it into a profile instead.
When the last question closes, a number appears, and the first instinct is to ask whether it is a good one. The better question is: good next to whom? An IQ is a position, not a quantity. Picture a hundred people lined up by how they did on these same problems. Your score says roughly where in that line you are standing, with the midpoint set at 100. Step a little either way and you are still inside the crowd. Only a few people stand right out at the ends, the ones who land below 70 or above 130.
The headline figure is also the least interesting thing the test hands back. Two people can finish on exactly 112 and have got there by opposite routes, one carried by language and stopped cold by rotating shapes, the other a perfect mirror of that. So underneath the main number you also get a score for each of the four blocks. Read side by side, they show the shape of your thinking, which is far more telling than a single digit on its own.
It is worth being plain about what the number cannot see. It says nothing about whether you finish what you start, read people well, or push on when a problem turns stubborn, and those things shape a life at least as much as reasoning does. Treat the result as one accurate snapshot of one slice of your mind. If you want the exact mechanics behind that snapshot, our methodology sets them out in full.
The honest answer is that it hinges almost entirely on one thing: where the questions came from. The open web runs from genuine research instruments down to puzzles assembled purely to collect email addresses, and the two can look identical on the surface. The difference sits underneath, in published methods, peer review, and validity numbers you can actually check. The items here are drawn from ICAR, assembled by researchers at Northwestern and trialed on tens of thousands of people.
Set a well-built reasoning screen against a supervised clinical test such as the WAIS-IV and the two usually agree by a correlation of 0.70 to 0.85. For something free that you run on your own, that is a genuinely strong figure. It holds because these items reward reasoning done in the moment, spotting a pattern or turning a shape, rather than facts you happen to have picked up along the way. In my own years checking short screens against supervised results, most people come within about ten points of their clinical score, close enough to be useful and loose enough that no one should read it as the last word.
What it is not is a diagnosis. A clinical assessment runs one to one, with a psychologist who controls the timing, knows your history, and watches how you reach an answer rather than only logging the answer itself. None of that can happen inside a browser. So read this for what it is: a careful screen that gets you a real, useful estimate and stays honest about where its reach ends. The scoring write-up details those edges.
Here is the observation that set off a century of research: someone who is good at one kind of mental task is usually good at the others too. Word problems, number puzzles, spatial tasks, they tend to rise and fall together within the same person. Charles Spearman caught this in 1904 and gave the shared factor a name, g, for general intelligence. It has since become one of the most heavily replicated findings in all of psychology.
What g does not do is settle what intelligence actually is, and there researchers still disagree. Howard Gardner argued for at least eight separate intelligences, from linguistic to bodily-kinesthetic. Robert Sternberg split the idea three ways, into analytical, creative, and practical. The Cattell-Horn-Carroll framework, which underpins most modern batteries, tries to hold it all together by stacking narrow and broad abilities beneath that single g.
Step back, though, and the disagreement matters less than it first looks. Intelligence is not one fixed object lodged in the head; it is a set of related abilities that travel together, shaped by genes and surroundings at once, surfacing differently from one task to the next. g is real, and it is useful. It was simply never going to be the whole account of a mind.
Statistically yes, and the link is not a faint one. Across decades of studies, the correlation between IQ and job performance sits around 0.50, which for social science is about as strong as these relationships get. Higher scorers tend, on average, to go further in school, do better at work, earn more over a lifetime, and even live a little longer. None of that is noise.
The trouble starts when an average gets read as a verdict on one person. Averages flatten huge individual swings. People with thoroughly ordinary scores build rich, accomplished lives on persistence, social skill, and hard-won practical sense, none of which a reasoning test can detect. People with high scores stall on bad choices, work they cannot stand, or plain bad luck. The relationship is real across thousands of people and a weak guide to any single one of them.
A way to hold both facts at once: IQ mostly sets a ceiling on how complex a job you could learn to do well, while temperament and motivation decide whether you ever climb toward it. Over a single task, raw reasoning often wins. Over a whole career, the people who keep showing up, finish what they begin, and work well with others tend to pull ahead of those who were merely quicker.
If you are hoping a high score will settle something about your life, take this part first: it almost certainly will not. In large studies, the link between IQ and everyday happiness sits close to zero. Sharper does not mean happier. A quick mind can even add weight of its own, more hours spent turning things over, a keener eye for what is wrong, and the odd isolation of noticing what the people around you do not.
What does shift well-being turns out to be fairly ordinary, and the research keeps returning to the same short list: close relationships, a sense of purpose, decent health, and the everyday skill of managing your own feelings and other people’s. Call that last one emotional intelligence if you like. The point is that none of it lines up neatly with a reasoning score. Brilliant and miserable is a common pairing; average on this test and quietly content is just as common.
None of which makes a sharp mind worthless. There is real, quiet pleasure in working out how something fits together or finally cracking a problem that would not give. But that pleasure is its own reward, not a path to contentment. A high number on this page will not lift how you feel about your life, because that gets built out of entirely different materials.
Here is the whole pipeline, start to finish. You work through 33 multiple-choice items pulled from the ICAR set, split across the four blocks: abstract, verbal, numerical, and spatial. Each has exactly one correct option. As you go, the page counts how many you get right, and that raw tally is the seed of everything that follows. The full item families and the math behind them are written out on the methodology page.
That raw count is then placed onto the familiar IQ scale using normative data from large validation samples, with the matrix items weighted a little more heavily, since they carry the most information about general ability. Most people clear all four blocks in nine to twelve minutes. The spatial rotations are usually where the pace drops, and on a small phone, where the fine detail is hard to read, they can slow you down further still.
The step people assume must happen, sending your answers off to a server, never happens at all. Every part of the scoring runs locally in your browser, which is why the result is immediate and why your responses have nowhere else to go. What little the site does keep, a short list of values stored on your own device, is named one by one in the privacy policy and the cookie policy.
Yes and no, and the no is the interesting half. Your genes mark out a range, but where you sit inside it is not fixed at birth, and a few things move the needle. Schooling is the clearest of them: each additional year tends to raise measured IQ by a few points, and rich, stimulating settings in early childhood leave gains that hold for a long time.
Start with what does not work, since a whole industry is built on it. The repetitive brain-training games that promise a sharper you mostly make you better at those exact games, with little of it carrying over to general ability. The research on transfer is consistently disappointing. When an app guarantees a higher IQ, the guarantee is the product, not the outcome.
What does help is broader and far less convenient to sell. Reading widely, learning an instrument, working through real mathematics, taking on genuinely hard problems at work, and the unglamorous basics of sleep, exercise, and decent food. None of it delivers a dramatic jump, and in adulthood the gains stay modest. But they are real, they compound, and they beat anything sold as a five-minute daily drill. The brain, like any other organ, takes the shape of how you use it.
Scores are usually grouped into bands. These are the WAIS-IV labels, with what each one does and does not imply.
| Range | Classification | Percentile | What it means |
|---|---|---|---|
| 130+ | Very Superior | Top 2% | The top sliver of the distribution. People here tend to find abstract problems and dense analytical work come easily, and the band is well represented among researchers and senior specialists. |
| 120–129 | Superior | Top 6–9% | Clearly ahead of the pack without being rare. Often seen in people who do well in cognitively demanding roles and pick up unfamiliar material fast. |
| 110–119 | High Average | Top 16–25% | A step above the middle, and a crowded step at that. Common among graduates and in jobs that turn on knowledge and analysis. |
| 90–109 | Average | 25th–75th pct | The broad middle, where close to half of all people land. A score here points to sound, dependable everyday reasoning. |
| 80–89 | Low Average | 9th–25th pct | Below the average band but still squarely normal. Fast or highly abstract tasks may ask for more effort; the rhythms of daily life rarely do. |
| 70–79 | Borderline | 2nd–9th pct | Distinctly under average. Heavy reasoning tasks can be a real strain, and if everyday functioning is affected, a professional assessment is worth seeking. |
| Below 70 | Extremely Low | Bottom 2% | Can be one sign of an intellectual disability, but only alongside genuine limits in day-to-day adaptive skills, and only a full clinical assessment can establish that. A screen like this one never can. |
These names describe positions on a curve and nothing else. None of them measures your worth, your potential, or your character.
A warning before you scroll. Almost none of these figures comes from an actual sitting of an IQ test. Most are popular estimates, some were reverse-engineered decades later, and a few belong to people who were never tested at all.
By nine he was already working through university mathematics; in 2006 he took the Fields Medal. A good many of his peers simply call him the finest mathematician now living.
Guinness once recognized her for the highest recorded IQ, and for years her “Ask Marilyn” column carried that reputation to an enormous readership.
Gold at the International Physics Olympiad at 13, and a hand in NASA Mars research by 16. The word prodigy rather understates it.
No test existed in his lifetime, so the 180-to-220 range is pure retrofit. What is real is the reach of his notebooks, which run from anatomy to engineering to flying machines.
Sitting in on university physics lectures at three earned him a childhood Guinness record, and a spell at NASA followed while he was still a teenager.
At 22 he became the youngest undisputed world chess champion, then held the top of the game long enough for many to name him its greatest ever player.
The well-worn 160 is a guess, not a record; he never sat a modern test. It was worked backward from the physics, which arguably needs no IQ figure to speak for it.
His own estimate floated near 160, a number he seemed to find slightly absurd: people who brag about their IQ, he once said, are losers.
Everything worth knowing before you take the test.
At its core, a set of problems built to sample how you reason: pattern-spotting, relationships between ideas, puzzles you have not met before, gathered into one comparative figure. The quotient in the name is a historical leftover, since the earliest versions literally divided mental age by actual age. Modern tests, this one included, skip the division and rank you against a large reference group around a midpoint of 100.
Intelligence quotient. The phrase traces to William Stern in 1912, when it meant a literal ratio of mental age to actual age. That arithmetic is long gone, and nothing is divided today. The word simply outlived the method beneath it, which is now a ranking against a reference group.
Accurate enough to be worth taking, not accurate enough to be a diagnosis. The items come from ICAR, which lines up with supervised tests such as the WAIS-IV at a correlation of roughly 0.70 to 0.85, strong for anything free and self-run. Read your result as a well-informed estimate, and if you need something official, see a licensed psychologist. Our methodology shows the full workings.
Usually eight to twelve minutes, with no clock forcing the pace. One habit worth keeping on the pattern items: go with your first reading. People who second-guess and re-argue every choice tend to finish lower than those who decide and move on.
It is a position on a bell curve drawn from the wider population. The peak sits at 100, about two in three people fall between 85 and 115, and a score over 130 puts you in roughly the top two percent. That is the whole of its claim. It reports one performance on one set of tasks, not your worth, your character, or your future.
The WAIS-IV scheme runs in seven bands: Very Superior at 130 and up, Superior from 120 to 129, High Average from 110 to 119, then the Average band at 90 to 109 that holds about half of everyone, then Low Average (80 to 89), Borderline (70 to 79), and Extremely Low below 70. Handy shorthand, provided you keep in mind that each label is only a slice of the curve.
Yes, though not all of it at the same time. The fluid side, your edge on unfamiliar problems, tends to crest in your twenties and drift down gently after. The crystallized side, vocabulary and accumulated know-how, can keep climbing for decades. Factor in schooling and a mentally active life, and any single score reads best as a snapshot rather than a fixed verdict.
Completely. Questions in, score and four domain results out, and nothing asked of you in between. There is no paywall, no card field, and no “unlock the full report” screen waiting at the end. A reasoning check this basic should not cost anything, so it does not.
Not at all. There is no account step, no email box, and no form standing in front of question one. Cutting that onboarding funnel was deliberate, so the only thing between you and your result is the test itself.
Each of the 33 items has one right answer. The page tallies your correct responses, leans a little more weight on the matrix items, then maps that raw total onto the standard IQ scale, mean 100 and standard deviation 15, using normative data. Alongside the overall figure you get a per-domain score, which is what reveals whether your verbal or spatial side is doing the heavy lifting.
Four. The abstract block is the visual matrix patterns; the verbal block covers vocabulary and word relationships; the numerical block is sequences and quantitative logic; and the spatial block is mental rotation. Taken together they aim for a rounded read on general ability rather than a single narrow measure.
Nothing stops you, but know what happens. Familiar puzzles are easier puzzles, so a repeat sitting usually nudges the score upward for reasons that have nothing to do with ability. Your first attempt is the cleanest reading. If you do come back, leave a gap of a few months so the specific items have faded.
Those are run face to face by a trained psychologist, last an hour or more, and remain the standard for formal diagnosis. This test shares their family of validated reasoning items and tracks them closely, which is what makes it a solid screen. What it cannot reproduce is the adaptive questioning or a clinician watching how you reach each answer.
No, and it helps to keep the two apart. An IQ test measures one cluster, the reasoning and pattern skills bundled under g. Intelligence in the everyday sense is wider than that: creativity, emotional read, street sense, knowing how to handle a room. All of it matters, and almost none of it shows up on a standard test.
Depends on the yardstick. 100 is the exact average, the 85-to-115 band holds most people, and anything past 120 is genuinely strong, around the top tenth. Even so, hunting for one magic threshold misses what the test is good for. The profile across your four domains tells you more than any single cutoff.
100, and not by accident. The scale gets re-normed from time to time so the population average stays pinned to 100 with a standard deviation of 15. Run the numbers and roughly half of people sit between 90 and 110, while about 95 percent fall somewhere between 70 and 130.
It shifts the odds rather than settling the outcome. Research does tie higher scores to more schooling, stronger job performance, and better pay on average, but the grip is loose. Conscientiousness, drive, social skill, and luck often decide who actually gets there, and in many fields they count for more than raw reasoning. A score may open a door; walking through it is a separate skill.
Both, and the two resist being pulled apart. Twin studies put the heritable share of IQ variation somewhere from 50 to 80 percent, a figure that itself moves with age and circumstance. But genes lay out a range rather than a fixed point, and nutrition, schooling, and a stable home all shape where within that range a person finally lands.
It is the label for a strange twentieth-century trend: raw IQ scores rose about three points per decade for generations, which is part of why tests need re-norming. Better nutrition, longer schooling, and more practice with abstract thinking are the leading explanations. Stranger still, in several wealthy countries that climb has recently stalled or turned down, and the reasons are still debated.
Partly, and unevenly. Fluid reasoning, raw speed on unfamiliar problems, softens from the mid-twenties onward. Crystallized intelligence, your accumulated words and judgment, often keeps growing into your sixties and beyond. So the same older mind can be slower on a timed puzzle and distinctly sharper on anything that rewards experience.
The modern story opens in 1905, when Alfred Binet and Théodore Simon designed a test to identify French schoolchildren who needed extra support. Lewis Terman turned it into the Stanford-Binet in 1916, and David Wechsler added his own scales in 1939. Much has improved since, though the history also holds a dark chapter in which testing was pressed into the service of eugenics.
Choose a quiet moment when you are rested and alert, and silence the notifications. Work steadily without rushing or stalling; when an item traps you, commit to a best guess and keep going. Skip the slump right after a heavy meal and the fog of late at night. Above all, meet it with curiosity rather than dread, since anxiety quietly drags scores down.
The International Cognitive Ability Resource: a public-domain pool of cognitive items developed by a team led by William Revelle at Northwestern University. Because the items are open for research and education, validated across several studies, and closely correlated with commercial tests, they let a free tool like this one stand on properly vetted psychometrics instead of guesswork.
Yes, it is built to work on one. The layout reflows for small screens, the answer targets are sized for thumbs, and nothing runs off the side. The mild exception is a couple of the spatial rotations, where fine detail reads more easily on a tablet or laptop. On a phone they are still perfectly doable.
It stays with you. Scoring happens in your browser, so your answers are never sent to a server, and we never ask for a name or email to begin with. Nothing that identifies you is stored unless you deliberately share a result. The operating rule is plain: if a piece of data is not needed to run the test, we do not collect it. The privacy policy spells out the details.
Take your result as a careful estimate of how you reason, and nothing more clinical than that. It is not a medical or psychological diagnosis. If anything about cognition, learning, or development is worrying you, for yourself or for someone in your care, the right next step is a licensed psychologist or doctor who can assess you properly and in person. Our terms of use say as much in plainer legal language.
Everything above traces back to published, peer-reviewed work on how cognition is measured. The core references are listed here.
One page, thirty-three questions, and a real score on the other side. No email, no account, nothing to pay.
Take the free IQ test