SAT to IQ converter

The SAT is not an IQ test, so no conversion is exact. These figures come from a statistical model: the SAT tracks general intelligence closely (about .8 in a large US study) but not perfectly, and high scores regress towards the average. If the SAT were a flawless IQ measure, a 1400 would mean IQ 123. It is not, so the honest estimate is lower and comes with a range.

Every SAT score, from the same model as the calculator. The likely range covers about four in five people at that score. The last column is the IQ you would get if the SAT measured IQ perfectly, which is the number most other charts show.

Going the other way, this is the SAT score people at each IQ typically get. Scores at the same IQ vary a lot, so the range matters more than the middle figure. Full IQ to SAT chart.

The SAT tests reading, writing and maths as taught in US schools. It is an achievement test on the surface, but achievement tests that sample many skills end up measuring general ability as well, because people who learn quickly do well across all the sections.

That overlap has been measured. Frey and Detterman (2004) compared SAT scores with a general-ability score built from the ASVAB military battery for 917 people from a national US sample. The correlation was .82. In a second study of 103 university students, the SAT correlated .48 with Raven's Advanced Progressive Matrices, or .72 after correcting for the fact that university students are a narrow slice of ability.

Two cautions apply. Those studies used the SAT as it was before 1995, and the "IQ" was a general-ability score from the ASVAB, not a Wechsler test. Nobody has published a large modern comparison of the current SAT with a Wechsler scale, so every SAT to IQ converter, this one included, is extrapolating.

Many converters simply line up percentiles: the SAT's top 1% becomes IQ 135 or so, and a near-perfect score becomes 145 to 160. That answers "how rare is this score?" It does not answer "how smart is this person?", because the SAT and IQ are only partly the same thing.

When two measures correlate at about .75, a person who is far above average on one is usually above average on the other but not as far. Statisticians call this regression to the mean. It is why the tall parents of tall children still tend to have children slightly closer to average height. Applied here, a 1600 is very rare (fewer than 1 in 100 test-takers), yet the most likely IQ behind it is about 131, with people scoring 1600 plausibly anywhere from 119 to 143.

The marker on the chart above shows the other reading. If the SAT measured IQ perfectly, a 1600 would equal IQ 140. The truth sits between the two, which is why a range is more useful than a point.

The scale has moved more than once. The SAT was recentred in 1995, which raised typical scores by roughly 100 points, and it was redesigned again in 2016 to the 400 to 1600 scale used now. The digital SAT, introduced in the US in March 2024, kept the 400 to 1600 scale but made each section adaptive: your performance on the first module decides how hard the second one is.

This converter uses the current scale and the Class of 2025 norms. If your score is from before 1995, a 1300 then is not a 1300 now, and pre-1995 charts (including the numbers in the research above) should not be read against today's table. College Board has published concordance tables between older and current scales if you need to translate an older score first.

Mensa is a useful reference point here. American Mensa accepted an SAT score of 1300 before September 1974 and 1250 from September 1974 to January 1994. It accepts no SAT scores taken after 31 January 1994, so an SAT cannot qualify you for Mensa today.

Read these as a guide to how your score compares with other test-takers, not as a verdict on intelligence. The estimated IQ column shows the most likely value, with the likely range in brackets.

There is no official SAT to IQ conversion. College Board does not publish one. What we can do honestly is treat the SAT as a good but imperfect signal of general ability. We place your score among SAT test-takers, assume how able that group is on average, and use the correlation between the two to predict the most likely IQ. Because the correlation is below 1, high and low scores are pulled towards the middle. That is why the estimate comes with a range.

Evidence grade: moderate. The SAT is one of the few tests here with a published comparison to a general-ability measure. It is still a comparison with a military battery, on the pre-1995 SAT, not with a Wechsler test on today's digital SAT.

Published facts we used

• SAT and ASVAB-derived general ability: r = .82 (N = 917), and .86 allowing for a curved relationship (Frey and Detterman, 2004).
• SAT and Raven's Advanced Progressive Matrices: r = .48, or .72 corrected for restricted range (N = 103).
• Class of 2025 mean 1029, SD 235, about 2.0 million test-takers (College Board annual report).
• Percentile table for actual test-takers (College Board score interpretation pages).

Assumptions we made

• Mean IQ of SAT test-takers: 102. Roughly half of US graduates now take the test, and school-day testing has brought the pool close to the general population. This is our assumption.
• IQ spread among test-takers: SD 14, slightly under 15 because test-takers are a selected group.
• Correlation between SAT and IQ among test-takers: .75, chosen below the published .82 because selected groups show weaker correlations and the studies used an older SAT.