The Highest IQ
Ever Recorded

1 The Highest IQ Ever: The Short Answer

Updated July 9, 2026 by Structural. There is no reliable "highest IQ ever," and the honest reason is that the number cannot be measured that high. Modern, properly normed intelligence tests top out at a Full Scale IQ of around 160, because scores beyond that are too rare to calibrate against real people. So when you read that someone has an IQ of 228, 250, or 300, you are not looking at a validated test result. You are looking at a different kind of number entirely.

It is worth separating two questions that get tangled together. One is whether some people are vastly more intelligent than others, which is clearly true. The other is whether we can put a precise, trustworthy number on the most extreme minds and rank them, which is not. This page is entirely about the second question. Nothing here doubts that geniuses exist; it doubts that a single figure like 228 or 300 can honestly capture or compare them, and that doubt turns out to be well founded.

Almost every famous sky-high IQ comes from one of a few sources: an old childhood ratio IQ that inflates for young prodigies, an estimate made by historians about someone long dead, a score from a non-standardized "high-range" test that no psychologist validated, or simply a figure a person or a fan attached to a name. None of these is comparable to the IQ you would get from a proper assessment today. This is exactly why Guinness World Records retired its "Highest IQ" category in 1990, after scientists objected that the measurement was not reliable at that level.

So this page is less a ranking than an honest tour of why the ranking cannot really exist, along with a fair look at the famous names attached to it. For the scale these numbers live on, see What Is the Average IQ?, and for just how rare high scores are, try our IQ Rarity Calculator.

2 First, What an IQ Number Actually Is

To see why the huge numbers are suspect, you have to understand how a modern IQ score is built. Today's tests use a deviation IQ: your performance is compared to a large, representative sample of people your own age, and your score reflects how far above or below the average you land. The scale is set so the average is 100 and one standard deviation is 15 points, which is the framework explained in our guide to the standard deviation of 15.

It also helps to notice what a deviation IQ is not. It is not a count of how much you know, a percentage of correct answers, or a measure of some absolute quantity of intelligence. It is purely a position relative to other people, a rank expressed on a familiar scale. That relative nature is why the extremes are so fragile: a rank only means something when there are enough people around you to rank against, and at the very top there simply are not, so the position stops being well defined.

That design has a built-in consequence for the top end. Because scores follow a bell curve, each step further from 100 describes a rarer and rarer slice of people. An IQ of 130 is already around the top 2 percent, roughly 1 in 44. An IQ of 145 is about 1 in 750. By 160, four standard deviations above average, you are looking at something like 1 in 30,000. The further out you go, the fewer real people exist to anchor the number, and that scarcity is the quiet reason a validated score cannot climb forever.

This matters because a deviation IQ is only as trustworthy as the sample it is compared against. A test can confidently tell 100 from 130 because it has measured thousands of people in that range. It cannot confidently tell 190 from 210, because it has never had a representative group of such people to calibrate on, for the simple reason that they are too rare to assemble. Keep that idea in mind, because it explains every strange number that follows.

3 Why Real Tests Stop Around 160

The major professional tests all place their ceilings in the same neighborhood, and it is not a coincidence. The Wechsler Adult Intelligence Scale, in both the WAIS-IV and the newer WAIS-5, reports a Full Scale IQ up to 160. The earlier WAIS-III stopped at 155. The children's version, the WISC-V, sits in the same range. The Stanford-Binet 5 also produces standard scores up to 160, with a specialized extended scoring option for assessing gifted children that reaches somewhat higher.

There is a practical reason publishers set the ceiling where they do rather than chasing higher numbers. Extending a test reliably to, say, 180 would require finding and testing enough genuine 180-level scorers to build norms, and since such people are rarer than one in a million, assembling a representative group of them is effectively impossible. Faced with that, a responsible test does the honest thing and caps its reporting, rather than printing precise-looking numbers it cannot actually stand behind. The ceiling is a mark of rigor, not a limitation of ambition.

That shared ceiling of about 160 is exactly four standard deviations above the mean, and it marks roughly where reliable norming runs out. Test publishers gather enormous samples, but even a sample of several thousand people will contain only a tiny handful scoring near 160 and essentially nobody scoring far beyond it. Without real people at those levels, there is nothing to compare a would-be higher score against, so the responsible thing a test can do is stop and label the top as "160 or above."

There are specialized ways to push past this. The Stanford-Binet has historically been favored for profoundly gifted assessment because of its extended ceiling, and its older Form L-M, along with modern extended-norm tables for some tests, can generate numbers above 160. But every one of these methods relies on extrapolation, extending the curve beyond where real data exists, and extrapolated scores get shakier the higher they go. A 165 from extended norms is a reasonable estimate; a 190 is a mathematical guess dressed as a measurement.

4 The Statistics of the Extreme

It is worth making the impossibility concrete, because it is genuinely the crux of the whole topic. To validly measure a score, you need a representative sample of people who actually score there, so the test knows what that level of performance looks like. The bell curve tells you how many such people exist, and the numbers get brutal fast.

A helpful way to feel the scale is to imagine trying to hold a contest to find the tallest person alive. Height, like IQ, follows a bell curve, and at the far tail the differences become vanishingly rare and hard to verify. Now imagine claiming to rank the top ten people by a height difference of a millimeter each, using a tape measure that was only ever calibrated for ordinary heights. That is roughly the position of anyone claiming to distinguish an IQ of 250 from one of 270. The tool was never built for that range, and no careful measurement there is possible.

An IQ of 160 is about 1 in 30,000. A score of 170 is roughly 1 in 1.5 million. A score of 190 is on the order of 1 in a billion, and 200 is rarer still. To properly norm a test at an IQ of 190, you would need to test a representative sample containing many people who score 190, which means testing several billion people, more than have ever taken any test in history. It simply cannot be done, which means a "measured" IQ of 190 or 200 is not a measurement in the normal sense at all.

This is the honest heart of the matter. The problem with the highest IQ scores is not that the people are not brilliant; many of them clearly are. The problem is that the instrument does not exist to place a precise number that high, and no amount of confidence in the person can manufacture a valid measurement. Beyond about 160, the scale stops describing a real, calibrated rank and starts describing an extrapolation, an estimate, or a guess.

5 Ratio IQ: Where the Giant Numbers Come From

If validated tests cap near 160, where do 228 and 250 come from? The answer, for the childhood cases, is an older and very different method called the ratio IQ. Before deviation scoring became standard, IQ was calculated as mental age divided by chronological age, multiplied by 100. A ten-year-old who performed like a fifteen-year-old would get a ratio IQ of 150, and a young child performing far beyond their years could generate an enormous figure.

This is also why old records and modern scores cannot simply be compared, even when both are called IQ. The word stayed the same while the method underneath it changed completely, from a ratio of ages to a rank against peers. Reading a 1950s childhood ratio IQ as if it were a modern deviation score is like reading a temperature in Fahrenheit as if it were Celsius because both use the word degrees. The label matches; the underlying quantity does not, and treating them as interchangeable is the root of most confusion about record IQs.

That formula behaves wildly at young ages. Because the denominator is a small chronological age, a prodigious child can produce a ratio IQ of 200 or more, not because that number means the same thing as a modern deviation IQ of 200, but because the arithmetic allows it. Ratio IQ was never designed to be comparable across ages, and it inflates dramatically for gifted children, which is precisely why the field abandoned it in favor of deviation scoring decades ago.

So the crucial translation is this: a childhood ratio IQ of 228 does not mean the person would score 228 on a modern adult test. It means a very bright child, tested young with an outdated method, landed on a number that method could produce. It is a real historical result, and it says something genuine about a precocious child, but it is not the same currency as the IQ scores we use today, and it cannot be compared to them directly.

6 Marilyn vos Savant and the Retired Record

The single most famous "highest IQ" belongs to Marilyn vos Savant, long cited at 228. The number is real in the sense that it came from a test: in 1956, at age ten, she took an adult-level Stanford-Binet and produced that score. Guinness World Records listed her for years as the person with the highest recorded IQ, and she became a household name for it, later writing a long-running magazine column.

It is worth being fair to vos Savant in both directions here. Critics sometimes use the ratio-IQ caveat to dismiss her, which goes too far; her later career, including decades of publicly answering difficult logic and probability questions, showed real and unusual ability. The point is narrower and applies to everyone: the specific figure of 228 should not be treated as a precise, modern measurement, regardless of how capable the person genuinely is. Respect for the person and skepticism about the number are not in conflict; they are both just honesty.

But her 228 is a textbook childhood ratio IQ, exactly the kind of number the previous section explained. It reflects a strikingly precocious ten-year-old measured with the older method, not a validated adult deviation score, and it cannot be lined up against a modern 160-ceiling test. This is not a knock on her; by any account she is exceptionally intelligent. It is a comment on the number, which carries all the caveats of ratio scoring at a young age.

The most telling part of her story is what Guinness did next. In 1990, Guinness retired the "Highest IQ" category altogether, after experts pointed out that IQ tests are not reliable enough at these extremes to crown a single record holder. The record was not handed to someone else; it was abolished as unmeasurable. When the official keeper of world records decides a category cannot be judged, that is the clearest possible signal about the whole idea of a highest IQ ever.

7 The Other Famous Names, Honestly

A handful of names circulate whenever the highest IQ is discussed. Each deserves a fair, specific note rather than a blanket number, because the sources vary and the caveats differ.

One reason these names recur is that a memorable number attaches to a person and then travels on its own, detached from where it came from. Once a figure like 228 or 210 is printed beside a name enough times, it hardens into a fact in the public mind, and the original context, a childhood test, an estimate, a media profile, quietly falls away. Restoring that context is most of what honest coverage can do. The figures are not lies so much as orphaned numbers that lost the caveats they were born with.

• William James Sidis is often given a wild range of 250 to 300. There is no validated test behind these figures at all; they are estimates and legend attached to a genuine early-twentieth-century prodigy. The numbers are essentially unsourced, and serious accounts treat them as unverifiable.
• Terence Tao, a Fields Medal-winning mathematician and one of the most gifted living scientists, is frequently cited around 220 to 230. Those figures trace to childhood testing and estimates, again in ratio-IQ territory. His genius is not in doubt; the specific number is not a validated adult score.
• Kim Ung-Yong, a Korean former child prodigy, was listed by Guinness with an IQ around 210, another childhood ratio figure from an era when such listings were still made.
• Christopher Langan, sometimes called "the smartest man in America," is quoted near 195 to 210. These come largely from media profiles and high-range testing rather than a standard clinical instrument, so they sit outside validated norms.

The pattern across all of them is consistent: remarkable people, real in their ability, paired with numbers that are either childhood ratio IQs, estimates, or scores from non-standard tests. Take the accomplishments seriously and the exact digits with a large grain of salt.

8 The Viral "Top IQ" Lists Are Mostly Fiction

If you have seen a graphic ranking the highest IQs in the world, with figures like 276, 300, or even 349 attached to names, it is worth knowing how little those lists mean. They circulate endlessly on social media, and they are assembled with almost no standards. A single list will freely mix childhood ratio IQs, historians' estimates of long-dead figures, self-reported scores from unregulated online or high-range tests, and numbers that appear to be invented outright.

It is also worth noticing what these lists never include: any source, method, or date for the numbers. A serious claim about a measurement would say which test was used, at what age, and under what conditions, because those details are what make a score interpretable. Their absence is not an oversight; it is the whole business model of the format, which trades on impressive-looking figures precisely because checking them is impossible. When a number arrives with no way to verify it, the safest assumption is that it cannot be verified because it never could be.

The tell is the size of the numbers themselves. As the statistics section showed, any figure much above 160 cannot be a validated measurement, so a confident claim of 276 or 349 is a giveaway that the number came from somewhere other than a properly normed test. Often these are scores people assigned themselves on untimed, take-at-home "high-range" tests that no independent body has validated, or they are promotional claims with no verifiable basis at all.

None of this means the people named are not intelligent, and it is not an accusation against any individual. It is a caution about the format. A ranked list of super-high IQs is entertainment, not data, and treating it as a factual leaderboard means trusting exactly the kind of number this entire page exists to question.

9 The Historical Geniuses: Estimates, Not Tests

Another whole class of high numbers belongs to people who died before IQ tests existed. You will see figures like 210 for Goethe, or estimates in the same range for Newton, Leibniz, or Einstein. These are not test results, obviously; they are historiometric estimates, produced by researchers who studied the recorded achievements, writings, and early development of eminent figures and worked backward to a plausible score.

There is a subtle circularity in these estimates worth flagging. A historian assigns a towering IQ to a figure largely because that figure achieved towering things, and then the number gets cited as if it explained the achievement. But the achievement was the evidence for the number in the first place, so the number adds no independent information; it just restates the person's eminence in a numeric costume. That does not make the estimates useless as a rough ranking of accomplishment, but it does mean they cannot tell you anything about intelligence that the person's actual work did not already show.

The best-known effort of this kind, from the 1920s, ranked historical geniuses by estimated childhood IQ based on biographical evidence. It is a genuinely interesting scholarly exercise, but it is estimation from achievement, not measurement of ability, and it inherits all the uncertainty of judging a person centuries later from what survived about them. We apply the same honest caveats when we discuss individual figures on pages like Einstein's IQ and Leonardo da Vinci's IQ, and gather them in our famous people and IQ hub.

So a historical genius "IQ" is best read as a shorthand for extraordinary achievement, not as a number anyone measured. It can be a fun and even defensible estimate, but it is a fundamentally different object from a score on a test, and stacking these estimates next to childhood ratio IQs and self-reported figures, as the viral lists do, compares things that were never the same kind of measurement.

10 High-Range Tests and IQ Societies

There is a small world of "high-range" tests, with names like the Mega Test, built specifically to try to discriminate among people at the extreme upper end where standard tests give up. They are usually untimed, taken at home over days or weeks, and heavy on very difficult puzzles. Some famous high numbers come from exactly these instruments, and high-IQ societies beyond Mensa sometimes use them for admission.

A further problem is that these tests measure a very specific and narrow thing: skill at extremely hard, mostly self-paced puzzles, often heavy on a particular style of reasoning. Even setting norms aside, that is not the same broad ability a full clinical battery estimates, so a high-range score and a Full Scale IQ are not just differently calibrated, they are partly measuring different things. Presenting one as if it were the other compounds the norming problem with a construct problem, and the resulting number tells you less than its size suggests.

These tests are a genuine attempt at a hard problem, and they are not worthless, but they are not standardized or validated the way professional tests are. Their norms are based on small, self-selected groups of people who chose to seek out a hard puzzle test, not representative samples, which is exactly the condition that makes a deviation score trustworthy. A "165" or "180" from such a test is not comparable to a clinically administered score, and the very self-selection that fills these tests biases their numbers upward.

The result is that scores from high-range tests occupy a strange middle ground: more serious than a random internet quiz, far less grounded than a professional assessment, and impossible to convert cleanly into the familiar scale. When one of these numbers appears on a "highest IQ" list, it is being treated as equivalent to a clinical score, which it is not.

11 So Who Really Has the Highest IQ?

Put all of this together and the honest answer to "who has the highest IQ ever" is that the question does not have a real answer, because the thing it asks for cannot be measured. There is no validated instrument that can rank the most extreme minds against each other, no representative sample to norm them, and no agreed method that would let you compare a childhood ratio IQ, a historiometric estimate, and a high-range test score on one scale. The leaderboard people want simply cannot be built from trustworthy data.

It is worth sitting with why this non-answer bothers people, because the discomfort is revealing. We are used to the idea that anything real can be measured and ranked, so the claim that the most extreme intelligence cannot be is faintly unsatisfying. But measurement has limits everywhere in science, and pushing past them produces false precision, not more knowledge. Accepting that the highest IQ is unmeasurable is not giving up on the question; it is answering it correctly, which is that honest measurement runs out well before the fantasy numbers begin.

What you can say is more modest and more true. A small number of people score at or near the top of what validated tests can measure, around 160, and that alone is extraordinarily rare. Beyond that ceiling, the differences between individuals become genuinely unmeasurable with current tools, so declaring one person the single "highest ever" is not a scientific statement but a story. The most intelligent people alive are real; a precise, verified ranking of them is not.

That is not a disappointing conclusion so much as a clarifying one. It frees you from taking any specific giant number at face value, and it redirects the interesting question from "what is the biggest number" to "what can we actually measure, and how rare is it." The second question has real answers, and they are more impressive than the fantasy leaderboard.

12 Putting Even "Normal High" Scores in Perspective

Because the conversation about record IQs is dominated by unmeasurable giants, it is easy to lose sight of how remarkable the scores tests can actually measure already are. You do not need a mythical 200 to be exceptionally rare.

There is also a healthier way to relate to your own curiosity here. Wondering where you stand is natural, and the good news is that the answer actually available, a validated score somewhere on the measurable part of the scale, is far more useful than a place on an imaginary leaderboard of giants. It comes with real norms, a meaningful percentile, and a profile of strengths, none of which a mythical 300 could ever offer. The measurable truth about yourself beats an unmeasurable fantasy about someone else every time.

Consider the real, calibrated end of the scale. An IQ of 130, the common threshold for the gifted range, is already rarer than 1 in 40. A score of 145 is about 1 in 750, putting such a person among a tiny fraction of any room they walk into. And 160, the ceiling of standard tests, is around 1 in 30,000, rare enough that most people will never knowingly meet someone who scores there. These are not mythical numbers; they are measurable, and they describe genuinely uncommon ability.

Seen this way, the obsession with 250s and 300s is a bit of a distraction from something more grounded and more interesting. The measurable top of the distribution is already astonishing, and unlike the record-book giants, it rests on real norms and real people. If you want to understand exceptional intelligence, the honest and fascinating place to look is the rare-but-measurable range, not the unverifiable numbers beyond it. Our rarity calculator lets you see exactly how uncommon any score is.

13 What This Means for You, and Where ACIS Fits

The practical lesson is simple. Treat any headline IQ above roughly 160 as a story rather than a measurement, ask where the number came from before believing it, and remember that the biggest figures are the least trustworthy precisely because they sit where no test can validly reach. Chasing a giant number, your own or anyone else's, is chasing something that cannot really exist.

A word on where our own test fits, in the same honest spirit as the rest of this page. ACIS gives you a real, validated score on the standard 100/15 scale, not a mythical one. It measures your reasoning across the cognitive domains and returns a full-scale score plus a domain profile, and its scale extends above the 160 ceiling of the standard Wechsler tests, into the mid-170s, so it can place unusually high scorers more finely than a test that stops at 160. But it does this honestly: the higher a score climbs, the more it should be read with humility, because the same statistical limits described here apply to every test, ours included. If you want to know where you actually stand, on real norms rather than a viral graphic, a validated test is the only way, and that is what ACIS is built to give you.

Sources and Further Reading

• On test ceilings: the Wechsler scales (WAIS-IV, WAIS-5, WISC-V) report a Full Scale IQ up to about 160, and the Stanford-Binet 5 uses the same standard range with an extended-scoring option for giftedness.
• Guinness World Records retired its "Highest IQ" category in 1990, citing the unreliability of IQ measurement at extreme levels.
• Marilyn vos Savant's cited 228 came from a childhood (age 10, 1956) adult-level Stanford-Binet, a ratio-IQ result rather than a modern deviation score.
• On ratio IQ versus deviation IQ and why the former inflates for young prodigies, see standard treatments of IQ scoring, and our guide to the standard deviation of 15.
• On historiometric estimates of historical figures, see our famous people and IQ hub and related pages such as Einstein's IQ.

14 FAQ: The Highest IQ Ever