The highest IQ ever officially recorded belongs to Marilyn vos Savant, whom Guinness World Records once listed at 228 — but that number comes from a childhood ratio-IQ score that modern testing cannot reproduce, and Guinness retired the category in 1990 for exactly that reason. Every figure you'll see above roughly 160 — for vos Savant, William James Sidis, Terence Tao or anyone else — is either an old ratio score, a retrospective estimate, or an unverified claim. Modern, properly normed IQ tests simply don't go that high. Here's what the famous numbers actually mean, and why the honest answer to "who has the highest IQ in the world" is: nobody can say.
Marilyn vos Savant and the Guinness record
Marilyn vos Savant entered the Guinness Book of World Records in 1985 under "Highest IQ" with a score of 228. The figure came from a Stanford-Binet test she took at age 10 in 1956, scored the old ratioway: mental age divided by chronological age, multiplied by 100. She answered questions at the level of an average 22-year-old's mental age while being 10 years old — hence roughly 228.
The problem is that ratio IQ and modern IQ are different measurements that happen to share a name. A ratio score of 228 does not mean she outscored 99.9999999% of adults; it means she was extraordinarily far ahead of other 10-year-olds. Modern tests use deviation scoring instead — your score reflects where you stand on the bell curve relative to your own age group, with a mean of 100 and a standard deviation of 15 (see our guide to the average IQfor how that works). Ratio and deviation scores aren't convertible, especially at the extremes. Guinness recognized this and retired the "Highest IQ" category entirely in 1990. Vos Savant remains a genuinely brilliant person — her long-running column famously handled the Monty Hall problem correctly while thousands of PhDs wrote in to disagree — but the 228 is an artifact of an obsolete scoring method, not a measurement any current test could produce.
William James Sidis: the 250–300 myth
William James Sidis (1898–1944) is the internet's favorite candidate for "smartest person ever," usually with an IQ quoted between 250 and 300. He was a real and remarkable prodigy: he entered Harvard at 11, lectured on four-dimensional bodies to the Harvard Mathematical Club, and reportedly worked with dozens of languages. But the IQ figure attached to his name is a retrospective estimate, made after his death and popularized by his sister and later writers — not a documented score from any standardized test. There is no verified record of Sidis ever sitting a proper IQ test. His story says a lot about prodigy, pressure and privacy; it says nothing measurable about the upper limit of human IQ.
Terence Tao, Kim Ung-yong and modern claims
Terence Tao, the Australian-American mathematician who won the Fields Medal at 31, is routinely cited with an IQ of 211–230. Again: this is an estimate, not a verified score from a normed test. What is documented is arguably more impressive — Tao scored 760 on the SAT math section at age 8 and remains one of the most productive mathematicians alive. Kim Ung-yong, a Korean former child prodigy, was listed by Guinness with an IQ of 210 under the same pre-1990 regime that listed vos Savant; the score reportedly derives from childhood testing and carries the same ratio-versus-deviation caveat. Kim himself has spent much of his adult life pushing back on the "failed genius" narrative, pointing out that a happy, ordinary career is not a tragedy.
Famous IQ claims: verified or not?
| Person | Claimed / estimated IQ | Basis | Verified? |
|---|---|---|---|
| Marilyn vos Savant | 228 | Childhood ratio score (Stanford-Binet, 1956) | Real score, obsolete scale — not comparable to modern IQ |
| William James Sidis | 250–300 | Posthumous retrospective estimate | No — no documented test score exists |
| Terence Tao | 211–230 | Estimate from childhood achievement | No — estimate only |
| Kim Ung-yong | 210 | Childhood score, pre-1990 Guinness listing | Same ratio-score caveat as vos Savant |
| Albert Einstein | 160+ (often quoted) | Pure speculation — he never took an IQ test | No |
| Stephen Hawking | 160 (often quoted) | Unsourced; Hawking said people who boast about IQ are "losers" | No |
Why modern IQ tests stop around 160
Modern tests like the WAIS-IV and Stanford-Binet 5 have measurement ceilings of roughly 160. This isn't timidity — it's statistics. Because IQ is a rank on the bell curve, extreme scores correspond to extreme rarity: an IQ of 160 is four standard deviations above the mean, roughly 1 person in 31,000. An IQ of 180 would be about 1 in 20 million; 195 would be around 1 in a billion. To norm a test at those levels you would need to find and test enormous samples of people who all score at the very top of every existing measure — which is practically impossible. Above the ceiling, the test also runs out of discriminating items: once someone answers essentially everything correctly, the instrument can't tell a 1-in-31,000 mind from a 1-in-20-million one. Our IQ percentile table shows how quickly rarity explodes past 145, and the IQ scale guide explains what the top classification bands actually mean.
Celebrity IQ lists are mostly fiction
Search any famous name plus "IQ" and you'll get a confident number. Almost none of these are sourced to a documented, professionally administered test. They propagate from listicle to listicle until repetition looks like evidence. Einstein never took an IQ test — the 160 attached to him is invented. The same goes for most figures quoted for actors, entrepreneurs and world leaders. When a celebrity score isreal, it's usually a childhood school test filtered through decades of retelling. A useful rule: if no one can name the test, the year and who administered it, the number is decoration.
What extremely high scores actually tell you
Within the measurable range, higher scores do predict real outcomes — academic achievement, job performance in complex roles — but the relationship is statistical, not a destiny, and past roughly 130 (the top ~2%, covered in what counts as a good IQ) differences in drive, opportunity and specific skills matter at least as much as further IQ points. It's also worth remembering that online IQ tests, including ours, produce estimates for self-knowledge and training — not clinical diagnoses, and certainly not world records. If a website tells you your IQ is 187, it is measuring your willingness to share the result, not your intelligence.
The honest bottom line: the highest verifiedscores on modern tests sit at the instruments' ceilings around 160, held by more people than you'd think and publicized by almost none of them. The mythic numbers — 228, 250, 300 — belong to a scoring system we abandoned decades ago, and to stories we like telling about genius.