Average IQ at MIT, Oxford and Yale: What the Scores Imply

What is the average IQ of MIT students? Nobody knows, because MIT has never published an IQ figure and does not test students for one. What MIT and Yale do publish is the SAT profile of each entering class, and that supports an estimate: about 129 to 134 for the middle student at either school, on the standard scale where 100 is average and the standard deviation is 15. Oxford is a different case, because it does not use the SAT at all. This piece applies the method from our Harvard IQ analysis to two more schools and says plainly where it stops working.

What do MIT and Yale actually publish?

Both schools file a Common Data Set each year, a standard report that includes the test scores of enrolled first-years. The 2025-26 editions, covering students who entered in fall 2025, report the following.

• MIT: SAT total at the 25th, 50th and 75th percentiles of 1520, 1550 and 1570; Reading and Writing 740, 760 and 780; Math 780, 790 and 800; ACT composite 34, 35 and 35
• MIT: 83% of enrolled first-years submitted an SAT (951 students) and 31% an ACT (353 students), with some sending both
• Yale: SAT total at the same three points of 1480, 1540 and 1560; Reading and Writing 730, 760 and 780; Math 740, 780 and 790; ACT composite 33, 34 and 35
• Yale: 67% of enrolled first-years submitted an SAT (1,092 students) and 25% an ACT (412 students); 91.4% of the SAT totals reported were 1400 or higher

Two cautions apply before any conversion. These are scores of enrolled students who chose to submit them, not of everyone admitted, and not of everyone who applied. And the national comparison matters: in the College Board's Class of 2025 report, the mean total score was 1029 with a standard deviation of 235, and 149,767 of about 2 million test takers, roughly 7%, scored 1400 or higher. At Yale, 91.4% of the SAT submitters did. MIT's official class profile page lists no test scores at all; the figures above come only from the Common Data Set.

How does an SAT score turn into an IQ estimate?

The link comes from a 2004 study by Meredith Frey and Douglas Detterman in Psychological Science, which found the SAT correlated .82 with a general cognitive ability factor in a national sample of 917 young Americans. Our explainer on SAT score to IQ conversion covers it in full. A correlation of .82 is strong but not perfect, so the prediction is pulled toward average, an effect called regression to the mean: multiply the student's standard-deviation position on the SAT by .82, then convert back to the IQ scale.

For a median student at the 99th percentile of a nationally representative sample, that is 2.33 standard deviations, which becomes 1.91 after the .82 factor and an IQ of about 129. At the 99.5th percentile it is about 132, and at the 99.7th about 134. Both schools' median section scores sit at or above the 99th percentile, and the College Board's top band is reported only as "99+", so the true rank inside that band is hidden. That is why the honest answer is a range, roughly 129 to 134, and why any one student's estimate carries a margin of about 12 points either way, the same margin derived in the Harvard piece.

Can the SAT rank MIT above Yale or Harvard?

No. The MIT median (1550) and the Yale median (1540) differ by 10 points, about 0.04 of the national standard deviation of 235. After the .82 regression that is roughly half an IQ point, far inside a margin of 12. Harvard reports section medians of 750 and 770 and no combined figure, so it lands in the same band. There is a second limit at the top: MIT's Math 25th percentile is 780 and its 75th is 800, the highest score the test gives, so at least a quarter of its SAT submitters hit the ceiling and the test cannot separate them from one another. Our piece on the IQ ceiling effect explains why every capped test behaves this way.

What about Oxford and Cambridge?

The SAT route does not exist there. Oxford and Cambridge select mainly on A-level or equivalent grades, admissions tests for many courses, and interviews, not on the SAT. Oxford is also replacing its own subject tests with three national ones (the TMUA, the ESAT and the TARA), starting with the October 2026 sittings for 2027 entry, according to tutoring-provider summaries; we could not open Oxford's own page to confirm the details. None of these tests is an IQ test, and neither university publishes an average admissions-test score or an IQ figure. We found no peer-reviewed IQ measurement of Oxford or Cambridge students either, so a figure for Oxford in the way this article gives one for MIT would be invented.

Why do claims of 150 or higher keep circulating?

Because the arithmetic is rarely done. To predict an average IQ of 150 (3.33 standard deviations above the mean) from the SAT, the students would need to sit about 4.06 standard deviations above the mean on the SAT, since 3.33 divided by .82 is 4.06. The SAT cannot register that: it tops out well before it. A school can be extremely selective, and Yale's 91.4% of submitters at 1400 or above shows that it is, without its students averaging anything close to 150. Selecting a class on test scores also compresses the range within it, which is why the middle-50% SAT bands above are so narrow. For how rare the high end really is, see how rare a high IQ score is.

What does a school average say about you?

Very little. A school's estimated average describes a group selected on many things, of which test scores are only one; our analysis of average IQ by education level and the average IQ by college major post show the same pattern of wide overlap between groups. To see where a specific score falls, the IQ percentile calculator turns any figure into a percentile, and our explainer on the IQ score margin of error shows why a single score is a range. If you want a baseline of your own, you can take the IQ Metrics IQ test.

The bottom line: the defensible answer to "average IQ at MIT, Oxford and Yale" is a range for two schools and a blank for the third. Roughly 129 to 134 for the middle student at MIT and Yale, unmeasured for Oxford, and no evidence that any of them averages 150.

Common questions