Is 125 a Good IQ Score? The Top 5% Explained

Haven't tested yet, or not sure your last result was reliable? You can take our 50-question assessment and get your score with a full percentile breakdown across four cognitive domains.

What does an IQ of 125 actually mean?

IQ scores are standardised so that the population mean is 100 and the standard deviation is 15. A score of 125 is 25 points above the mean, which is one and two-thirds standard deviations. That places you clearly inside the range psychologists describe as high intelligence, without being at either edge of it.

125 occupies an unusual position among the scores people look up. Most searched-for numbers are boundaries: 120 opens the Superior band, 130 opens the gifted range and qualifies for Mensa. 125 is not a boundary at all. It sits in the middle, which means nothing about it is an artefact of where someone drew a line on a chart. What you have is a clean, unambiguous position in the distribution.

• IQ 90-109: Average
• IQ 110-119: High Average
• IQ 120-129: Superior, and 125 sits squarely in the middle
• IQ 130 and above: Very Superior, conventionally called gifted

As with the neighbouring scores, the label has been modernised. The Wechsler family adopted value-neutral descriptors with the WISC-V in 2014 and carried them into the WAIS-5 in 2024, so a current clinical report calls 120-129 Very High rather than Superior. The ranges are unchanged.

The exact percentile for an IQ of 125

Using the standard normal distribution with a mean of 100 and a standard deviation of 15, an IQ of 125 maps to the 95.22nd percentile. Here is how it sits against its neighbours:

The 95th percentile is worth dwelling on because it is the most widely used landmark in applied statistics. It is the cut-off behind confidence intervals, significance testing and most “top 5 percent” claims in everyday writing. At 125 you are standing exactly on it.

In concrete terms: about five people in a 100-person office. Around 48 in a secondary school of 1,000. Roughly one carriage-worth on a busy commuter train. You will meet people at or above your level regularly, but in most small groups you will be at or near the top.

The most famous 125 in history

No other IQ score carries a single anecdote the way 125 does. Richard Feynman, Nobel laureate in physics and one of the most obviously brilliant people of the twentieth century, is reported to have scored exactly 125. The story is used constantly, usually to argue that IQ tests are worthless. It is a better story than that, and the details matter.

The source is James Gleick's biography Genius, which records a school IQ estimate of 125 and describes it as “merely respectable.” The test was a group test administered at his school in Far Rockaway, New York, when he was around 15. Its name, its contents and the original record have never surfaced. No professionally administered score for Feynman was ever documented. So 125 is accurately described as reported, not verified, and it rests on a single adolescent sitting.

Two features of tests like that one matter here. School group tests of the era were heavily weighted towards verbal content, and they had low ceilings, meaning they ran out of hard questions well before they ran out of capable students. Both properties would systematically understate someone whose strongest suit was mathematical and spatial reasoning. Feynman is close to the archetype of that profile.

He also enjoyed the story. He is widely said to have joked that his IQ was too low for Mensa, though no documentation of any such exchange has ever been published, so that part is best treated as a joke he liked rather than an event that happened. His sister Joan, herself a distinguished astrophysicist, scored 126 and reportedly never let him forget the extra point.

Here is the part almost every retelling leaves out. Feynman is an outlier in this data, not a representative case. When Anne Roe studied 64 eminent scientists in the early 1950s using a purpose-built high-ceiling test, her subjects scored extraordinarily highly, with estimates placing the typical eminent scientist far above 130. The exact conversion to a modern IQ scale is disputed, because Roe used a custom instrument, but the direction is not in question. The typical person doing world-class scientific work scores well above 125, not at it.

So the honest lesson of Feynman's 125 is narrower and more useful than “IQ doesn't matter.” It is that a single score from one imperfect test is weak evidence about one individual, particularly when the test was short, verbally weighted and taken at 15. That is a claim about measurement error, which is well established. It is not evidence that ability is irrelevant to achievement, which the broader data contradicts.

How much to trust a single score of 125

Every IQ test carries measurement error. The WAIS-IV has a test-retest reliability of about 0.96 for Full Scale IQ, which is excellent but not perfect, and the standard error of measurement is typically plus or minus 3 to 5 points. If you scored 125, your true underlying ability most likely sits between 120 and 130.

That band is worth looking at closely, because it spans from the bottom of the Superior range to the start of the gifted range. Your measured 125 is consistent with a true ability anywhere in that span. Retest on a poor day and 121 is plausible. Retest rested and unhurried and 129 is equally plausible. None of those numbers would represent a change in you.

Unlike a score sitting exactly on a band edge, though, a 125 does not flip classification under normal variation. You would need an unusually bad sitting to fall out of Superior and an unusually good one to clear 130. That stability is the quiet advantage of a mid-band score, and it means a 125 from a decent test is telling you something reasonably solid.

Can you join Mensa with an IQ of 125?

Not on that score. Mensa requires a score at or above the 98th percentile on a supervised standardised test, which is IQ 130 on the Wechsler scale and 132 on older Stanford-Binet scales using a standard deviation of 16. An IQ of 125 is at the 95th percentile, five points short.

But 125 is the first score in this series where that gap is genuinely worth taking seriously, and it is worth being precise about why. Five points is inside the standard error of measurement on most tests. Someone whose true ability is 128 will score anywhere from roughly 123 to 133 across sittings. If your 125 came from a single session, a supervised test on a different day could legitimately land above the line.

This is not an argument for retaking a test until you get the number you want. Repeated attempts inflate scores through practice effects, and a score obtained that way is not measuring what you wanted it to measure. The legitimate version of the argument is narrower: if you have specific reason to think one sitting understated you, such as illness, a rushed administration, or a short test with a low ceiling, then a single properly supervised retest is a more accurate estimate rather than a second roll of the dice.

Worth keeping in proportion, though. Mensa is a social club with a cognitive entry ticket, and membership correlates weakly with income, career progression, relationships or life satisfaction. At the 95th percentile you already have more cognitive headroom than any career realistically demands. Clearing the line would change your membership status and very little else.

IQ 125 and education

An IQ of 125 is comfortably above the average for every category of educational attainment in Kaufman and colleagues' (2016) analysis of WAIS-IV normative data, doctorates included. For comparison:

• High school graduates: mean IQ around 99
• Bachelor's degree holders: mean IQ around 110
• Master's degree holders: mean IQ around 112
• Professional degrees (MD, JD): mean IQ around 114
• Doctoral degrees (PhD): mean IQ around 115
• IQ 125: 10 points above the mean for PhD holders

Ten points above the doctoral mean is a substantial margin. It means that in a room of PhD holders in a mixed field, you would be expected to be in the upper portion of the room on this particular measure. Whether you have a doctorate is a separate question entirely, driven by interest, funding, timing and appetite for years of institutional process.

As always, the correlation between IQ and education mostly reflects selection rather than causation. More able people tend to pursue more education; education itself raises measured IQ only modestly, at roughly 1 to 5 points per additional year according to Ritchie and Tucker-Drob (2018). A 125 with no degree is exactly as real as a 125 with two.

What careers match an IQ of 125?

At 125 you are at or above the group mean for every profession that has been systematically measured, including the most quantitatively demanding ones. The occupational literature (Gottfredson 1997; Hauser 2002) places even the highest group means, in theoretical physics and mathematics research, in the 125-130 band. You are inside it.

• Research scientists across the physical and life sciences
• Mathematicians, statisticians and theoretical physicists
• Physicians and surgeons in every specialty
• Quantitative analysts, actuaries and research economists
• Software engineers in algorithmically demanding roles
• Barristers, appellate lawyers and legal academics
• University faculty across all disciplines
• Senior engineers, systems architects and R&D directors

Which means the useful question at 125 is no longer what you are capable of. It is what you are willing to become expert in. Past this point the differences between people in demanding fields are driven overwhelmingly by accumulated domain knowledge, sustained attention on hard problems, and the willingness to stay with something long after it stops being novel. Those are built, not issued.

What to keep in mind about an IQ of 125

A few things worth knowing so you can read your score accurately, not to talk it down, but because understanding them makes the number more useful.

• You are mid-band, which is a good place to be. Unlike 120 or 130, your score does not depend on where someone drew a classification line. Normal test-to-test variation will not move you out of the Superior range. Whatever a 125 tells you, it tells you fairly stably.
• The Feynman story cuts both ways. It is genuinely true that a single score is weak evidence about one person, and that should make you hold your own 125 loosely. It is not true that ability is irrelevant to achievement. Use the story as a caution about measurement, not as permission to dismiss the measurement entirely.
• The advantage is real and probabilistic. Cognitive ability predicts job performance, with correlations around 0.5 for complex roles, along with educational attainment and income. At the 95th percentile those odds run clearly in your favour. They remain odds. Conscientiousness, health, timing and network account for most of what is left.
• It captures one kind of intelligence. IQ measures the general factor behind reasoning and problem-solving, which is important and heavily studied. Creativity, judgement, emotional intelligence and deep domain expertise shape outcomes too, and a 125 tells you nothing about any of them. They could easily be higher.

The bottom line on IQ 125

An IQ of 125 is a strong score by any reading. You are at the 95th percentile, roughly 1 in 21, ten points above the average PhD holder, and at or above the group mean for every profession that has been measured, including theoretical research. You sit in the middle of the Superior range rather than on its edge, which makes the score unusually stable.

Two things are worth carrying away. First, the five points between you and the gifted threshold are inside normal measurement error, so if that line matters to you, a single supervised retest is a reasonable thing to do once, and a bad thing to do repeatedly. Second, the most famous 125 in history belonged to a physicist who reshaped his field, which should tell you that a number from one sitting of one test is a weak description of a person, in both directions.

And if your 125 came from a short quiz, it is worth checking against something with enough hard questions to measure the top of the range properly. That was the problem with Feynman's test, and it has not gone away.

Frequently asked questions