What Is a Good IQ Score? Percentiles, Averages, and What the Numbers Mean
A plain-English guide to IQ score ranges: what counts as average, above average, or gifted, how percentiles work, and why a single score doesn't tell the whole story.
You just took an IQ test, or you’re thinking about taking one, and the obvious question is sitting right there: what number counts as “good”? Is 100 fine? Is 115 impressive? Does anything under 130 even matter?
The honest answer is that “good” depends entirely on what you mean by it. IQ scores aren’t graded like a school test where 90% is an A. They’re built on statistics, specifically on where you land compared to everyone else who’s taken the same test. Once you understand that system, the number stops feeling mysterious and starts telling you something genuinely useful, as long as you also understand what it leaves out.
This guide walks through how IQ scoring actually works, what the common score bands mean, how to read a percentile, and where the limits of a single number really are. If you want to see where you fall, you can take our free IQ test and use this article to interpret the result.
How IQ scoring actually works
Modern IQ tests are built around a statistical shape called the normal distribution, or bell curve. If you gave the same test to a huge, representative sample of people and plotted all their scores on a graph, you’d get a curve that’s tall in the middle and tapers off symmetrically on both sides. Most people cluster near the center, and progressively fewer people show up as you move toward either extreme.
Test makers deliberately design their scoring so that this curve has two fixed properties:
- The mean (average) score is set at 100.
- The standard deviation, a measure of how spread out scores are, is set at 15 points.
Standard deviation is the real workhorse here. It tells you how far a typical score sits from the average. Because IQ tests are normed this way, a score of 115 isn’t just “higher than 100,” it’s specifically one standard deviation above the mean. A score of 130 is two standard deviations above, and a score of 85 is one standard deviation below. This consistent spacing lets psychologists translate a raw number into a percentile, and it’s why the 100/15 model is the standard across most well-known tests, including the Wechsler scales and Raven’s Progressive Matrices, a widely used matrix-reasoning test that asks you to spot patterns in abstract shapes.
It’s worth pausing on that word “normed.” A test’s scoring only makes sense in reference to the specific group used to build it: a large, carefully selected sample whose results become the yardstick everyone else is measured against. That’s why a 110 on one test and a 110 on another aren’t automatically identical. Both tell you “10 points above this test’s own average,” but the tests themselves might weigh different skills or draw on different norming samples. A score only means something relative to the scale that produced it.
Turning a score into a percentile
A percentile answers a simple question: out of 100 people who took this test, how many scored at or below you? If you land in the 84th percentile, that means your score is higher than roughly 84 out of 100 test takers in the norming sample, and lower than about 16.
Because the bell curve is symmetrical and its shape is mathematically predictable, psychologists can calculate percentiles directly from standard deviations. Here’s the rough breakdown that falls out of the normal curve:
- About 68% of people fall within one standard deviation of the mean, meaning a score somewhere between 85 and 115.
- About 95% fall within two standard deviations, or roughly 70 to 130.
- Well over 99% fall within three standard deviations, or roughly 55 to 145.
This is sometimes called the 68-95-99.7 rule, and it holds for any normal distribution, not just IQ. It explains something that surprises people the first time they see it: scoring 100 doesn’t make you some exact midpoint of “smart,” it puts you at the 50th percentile, meaning half of test takers scored higher and half scored lower. A score of 115 isn’t a small bump either; it already puts you ahead of about 84% of people, because the density of scores thins out fast once you move away from the center.
IQ score ranges and what they mean
Different psychologists and test publishers use slightly different labels, but the descriptive bands below are close to the conventions used across most modern intelligence tests. Treat the percentile figures as approximate, since exact norming samples vary a bit from test to test.
| Score range | Approximate percentile | Common descriptive label |
|---|---|---|
| Below 70 | Below 2nd percentile | Extremely low |
| 70-84 | 2nd to 14th percentile | Borderline to low average |
| 85-114 | 16th to 84th percentile | Average |
| 115-129 | 84th to 97th percentile | Above average / high average |
| 130-144 | 97th to 99.6th percentile | Superior / very superior (often called “gifted” range) |
| 145+ | Above 99.6th percentile | Exceptionally gifted |
A few things worth unpacking about this table.
Average covers more ground than people expect
The single biggest misconception about IQ scoring is that “average” means “unremarkable” or, worse, “mediocre.” Look at the range again: 85 to 114 is classified as average, and it covers roughly two-thirds of the entire population. That’s not a narrow band squeezed around 100, it’s the majority of humanity. If you score 90 or 112, you’re both solidly inside the normal range of human cognitive variation, just on different sides of the midpoint. Neither number is an insult, and neither is a trophy; it’s simply where most people land, by definition.
The bands compress as you move outward
Notice how the ranges in the table get progressively narrower relative to how rare they are. Going from “average” to “above average” only requires crossing from 114 to 115, but the population represented by each subsequent 15-point band shrinks fast. That’s the nature of a bell curve: the tails are thin. A jump from 130 to 145 represents a much rarer group of people than a jump from 100 to 115, even though both are “15 points."
"Gifted” is a threshold, not a single definition
You’ll often see 130 cited as the cutoff for “gifted,” largely because many gifted-identification programs use two standard deviations above the mean as their benchmark. But this isn’t a universal rule. Some programs use 125, some use 132, and some rely on multiple criteria beyond a single test score, including teacher recommendations or performance on several different assessments. If you’re researching a gifted program, check that specific program’s own criteria rather than assuming 130 is a fixed standard.
So what actually counts as a “good” score?
Here’s the honest, slightly unsatisfying answer: it depends on what you’re using the score for.
If you’re just curious about where you stand relative to other people, then anything comfortably in the average range (85-114) is, statistically speaking, entirely normal and fine. There’s no threshold below which a score becomes a problem in daily life; the average range represents typical, healthy cognitive functioning for the large majority of the population.
If you’re applying to a specific program or assessment with a stated cutoff, then “good” simply means “meets or exceeds that cutoff.” That’s a practical, context-specific bar, not a universal one. And if you’re comparing your own results over time, “good” might just mean “higher than last time,” though that comparison needs some caution, which we’ll get to next.
A single adjective like “good” doesn’t map onto IQ scores the way it does onto a pass/fail measure. The bell curve model describes relative standing, not a verdict.
What an IQ score can’t tell you
This is the part that gets glossed over in a lot of pop-science coverage of IQ, and it deserves real space here.
It measures specific types of reasoning, not “intelligence” as a whole
Most IQ tests are built around a cluster of skills: pattern recognition, working memory, processing speed, spatial reasoning, and verbal comprehension. These are genuinely useful abilities, and they correlate with things like academic performance. But they aren’t the same as creativity, social skill, practical know-how, emotional intelligence, or the judgment that helps someone navigate a messy real-world situation. Someone can score modestly on a matrix-reasoning test and still be an exceptional negotiator, mechanic, writer, or parent. The test simply wasn’t built to measure those things.
Scores can shift with conditions, practice, and age
A single test result is a snapshot, not a permanent label. Fatigue, stress, illness, unfamiliarity with the test format, and even the room you’re sitting in can all nudge a score up or down on a given day. Scores can also shift somewhat with practice: someone who has taken several matrix-reasoning tests before will generally perform a bit better on their next one, simply from familiarity with the format rather than a genuine change in reasoning ability. Cognitive performance on certain subtests also naturally changes across the lifespan, which is part of why professionally administered tests use age-adjusted norms rather than one flat scale for everyone.
There’s also a well-documented historical pattern called the Flynn effect: average raw scores on IQ tests have generally risen across generations over the course of the 20th century, which is one reason publishers periodically re-norm their tests. This doesn’t mean any one person’s underlying ability changed; it reflects shifts in the broader population over time, and it’s a reminder that the “average” of 100 is tied to a particular norming sample and era, not a fixed law of nature.
Different tests, different scales
Because norming samples and question sets vary, a score from one test isn’t automatically interchangeable with a score from another. A well-constructed test with a large, representative norming sample gives a more reliable number than a short, informal quiz, but that doesn’t mean quick online tests are worthless. They’re a fun, low-stakes way to explore your own reasoning strengths, especially when the format is transparent about how it estimates a score. Just don’t treat a casual result as clinically equivalent to a full psychological assessment. If you want a quick, well-structured starting point, take the test yourself and treat the result as a data point rather than a diagnosis.
It helps to separate the score from your self-worth entirely. Cognitive tests measure a narrow slice of mental performance under timed, artificial conditions. They’re useful for research, some clinical decisions, and satisfying your own curiosity, but they’re a poor tool for deciding what you’re capable of or how valuable you are as a person. Plenty of people with modest test scores build extraordinary careers, and plenty of people with sky-high scores struggle with things a test never measured in the first place, like discipline or communication.
Frequently asked questions
What is considered a “genius” IQ score?
There’s no official cutoff for “genius,” and the word isn’t a formal psychometric category. Some sources informally use 140 or higher, corresponding to roughly the 99.6th percentile or above, but genuine genius-level achievement in any field usually involves ability, creativity, and sustained effort that no single test score captures.
Is 100 a bad IQ score?
No. A score of 100 is, by definition, the statistical average of the norming population, sitting at the 50th percentile. It represents perfectly typical cognitive performance and is neither low nor high, just the midpoint of a very wide normal range.
Can you actually raise your IQ score?
Scores can shift somewhat with practice on similar question formats, better test conditions, or improved focus and sleep, and certain training may sharpen specific skills like working memory. But a large, stable jump in underlying reasoning ability isn’t something that happens quickly, and repeated attempts at the same test aren’t always perfectly comparable.
Why do I get a different score on different tests?
Tests vary in the specific skills they emphasize, the length and difficulty of their question sets, and the norming sample used to calibrate their scale. A score is only meaningful relative to the test that produced it, so some variation between tests, especially between a short informal quiz and a longer standardized assessment, is completely expected.
Does a high IQ score guarantee success?
Not on its own. IQ scores correlate with certain academic and cognitive outcomes, but real-world success also depends heavily on factors a reasoning test doesn’t measure, including motivation, communication skill, emotional regulation, creativity, and plain persistence.
What’s the difference between IQ and emotional intelligence?
IQ tests focus on measurable reasoning tasks like pattern recognition, memory, and verbal or spatial problem-solving. Emotional intelligence refers to a different set of skills entirely, things like reading social cues, managing your own reactions, and navigating relationships. The two are assessed in completely different ways, and a strong score in one doesn’t predict a strong score in the other.
