The average IQ is 100 at every age, by design. Modern IQ tests are age-normed: your answers are compared only with a large sample of people your own age, and the scores are scaled so that the middle of each age group lands at 100, with a standard deviation of 15. So the average IQ for a 13-year-old is 100, the average IQ for adults is 100, and the average for people in their 70s is also 100.
What does change with age is raw ability: how many problems you can solve, how fast, and how much you know. That change is real and large. It just does not show up in the IQ number, because the yardstick moves with you.
Why IQ is always 100 on average
When a publisher builds a test such as the WISC-V (for children aged 6 to 16) or the WAIS-5 (for adults aged 16 to 90, released by Pearson in 2024), it tests a large sample chosen to match the population on age, sex, region, education and ethnicity. Each age band gets its own score table. For children the bands span only a few months, because abilities change quickly; for adults they span several years.
Your raw score is then converted to a standard score within your band. If you perform exactly like the median person of your age, you get 100. If you outperform about 84% of your age group, you get 115. That is why an IQ is best read as a position on the bell curve, not a quantity of intelligence. The conversion from raw score to IQ works the same way at every age.
Early tests did not work this way. Binet's original scale reported a mental age, and William Stern proposed dividing mental age by chronological age, and Lewis Terman multiplied the result by 100: a 10-year-old performing like a typical 12-year-old scored 120. That ratio breaks down for adults, since raw ability stops climbing in the late teens, so a 40-year-old would appear to lose IQ every birthday. David Wechsler replaced it with the deviation IQ described above, and every major modern test uses that approach. The history of IQ testing covers the switch in more detail.
Average IQ by age: the table
The left column is the same in every row. The right column is the part that actually moves.
| Age | Average IQ | What is changing in raw ability |
|---|---|---|
| 3–5 | 100 | Rapid growth in language, attention and memory; scores at this age are the least stable over time |
| 6–10 | 100 | Steady gains in reasoning, vocabulary and working memory; schooling starts to add knowledge each year |
| 11–13 | 100 | Abstract reasoning keeps improving; a typical 13-year-old solves noticeably harder problems than a typical 10-year-old |
| 14–17 | 100 | Raw scores on most reasoning and speed tasks approach adult levels |
| 18–29 | 100 | Processing speed and fluid reasoning are at or near their peak |
| 30–44 | 100 | Fluid reasoning and speed begin a slow decline; vocabulary and general knowledge are still growing |
| 45–64 | 100 | Crystallized knowledge is near its high point; speed and new-problem solving are lower than at 25 |
| 65–79 | 100 | Declines in speed, memory and reasoning become easier to notice; vocabulary is largely preserved |
| 80+ | 100 | Declines are steeper on average and differences between individuals are wide, often tied to health |
The age ranges are rough. Researchers disagree about exactly when each change begins, and the answer depends partly on study design, as explained below.
Average IQ for children and teenagers
A child's IQ compares them with children of the same age, often to the month. A 13-year-old with an IQ of 100 is reasoning, remembering and working at the speed of a typical 13-year-old. The same child would score well above 100 on the 10-year-old tables and below 100 on the 16-year-old tables.
This matters for parents who see a score and ask whether it is good for the age. The score already accounts for age. A 7-year-old scoring 115 and a 15-year-old scoring 115 sit at the same position relative to peers, roughly the 84th percentile, even though the 15-year-old answers far harder items.
Scores from early childhood are less predictive than later ones. Tests for preschoolers lean heavily on language and attention, which develop at different speeds in different children, and a single sitting is easily affected by mood or unfamiliarity with testing. Stability rises through the school years. In the classic summary by the American Psychological Association's task force, correlations between scores measured a few years apart are substantial from middle childhood onward, while infant measures predict later IQ only weakly (Neisser et al., 1996).
Teenagers can still change. In one well-known study, 33 adolescents were scanned and tested at ages 12 to 16 and again three to four years later. Some individuals' scores rose or fell by 15 points or more, and the changes tracked changes in brain structure (Ramsden et al., 2011). The sample was small, but it is a useful reminder that a teen's score is not fixed. Our article on whether IQ can change over time covers this in depth.
Average IQ for adults, and how ability changes after 20
Adult tests split ability into broad areas, and those areas age differently. The distinction goes back to Raymond Cattell and John Horn, who separated fluid intelligence (solving new problems, spotting patterns, holding information in mind) from crystallized intelligence (vocabulary, facts, learned procedures). In their early cross-sectional data, fluid ability was lower in each older age group while crystallized ability held steady or rose (Horn & Cattell, 1967).
Later work has refined the picture rather than overturned it:
- Processing speed peaks earliest, around the late teens or early 20s.
- Fluid reasoning and memory tend to peak in the 20s and decline gradually afterward. Timothy Salthouse, using large samples, concluded that some age-related decline in reasoning, speed and memory is detectable in healthy, educated adults by their 20s and 30s (Salthouse, 2009).
- Vocabulary and knowledge keep rising into middle age or later. In a study of tens of thousands of online participants, different abilities peaked at very different ages, with vocabulary peaking in the late 60s or early 70s (Hartshorne & Germine, 2015).
Because the WAIS-5 and similar tests norm each adult age band separately, a 70-year-old with an IQ of 100 has performed like a typical 70-year-old. A 25-year-old producing the same raw answers on the speed and reasoning subtests would score lower, because the 25-year-old comparison group sets a higher bar there.
Cross-sectional versus longitudinal findings
How early decline begins is still debated. Cross-sectional studies compare different people at different ages at one moment, and they tend to show earlier decline. But older cohorts also grew up with less schooling and different environments, which inflates apparent age differences; this overlaps with the Flynn effect. Longitudinal studies follow the same people over decades and usually show later, smaller declines, as in the Seattle Longitudinal Study, where many abilities held up fairly well until the 60s (Schaie, 2005). Longitudinal studies have their own bias: people who stay in a study for decades tend to be healthier, and taking the same tests repeatedly brings practice gains. The honest summary is that speed and fluid reasoning decline on average across adulthood, with the timing still contested, and the decline speeds up in later life.
Does your IQ stay the same as you age
Here two ideas need to be kept apart. Your raw abilities change a great deal across life. Your rank among people your age is remarkably stable once you are past early childhood.
The best evidence comes from Scotland. On June 1, 1932, nearly every 11-year-old in the country took the same mental ability test as part of the Scottish Mental Survey. Decades later, Ian Deary and colleagues retested survivors on the same test. Scores at age 11 and age 77 correlated at 0.63, rising to about 0.73 after adjusting for the restricted range of the people who were retested (Deary et al., 2000). A follow-up of the Lothian Birth Cohort 1921 found a correlation of about 0.54 between ages 11 and 90, around 0.67 after the same kind of adjustment (Deary et al., 2013).
Those are strong correlations over 65 to 80 years. They also leave real room for change: a correlation of 0.6 to 0.7 means many people moved noticeably up or down relative to their peers. Health, education, occupation and lifestyle are among the factors researchers link to how well individuals hold their rank in old age, though separating cause from effect is difficult.
Cognitive decline versus normal aging
Normal aging tends to look like slower recall of names, needing more time to learn a new app, or losing track when juggling several things at once. Vocabulary, judgment built on experience and long-practiced skills usually hold up. Changes that disrupt daily life, such as getting lost in familiar places, repeating the same questions, or trouble managing money or medication, are not part of typical aging and are a reason to see a doctor. An IQ score, online or otherwise, cannot diagnose dementia or any other condition, and nothing in this article is medical advice.
What this means when you take a test
If a test reports an IQ, check that it was normed by age. An adult test without age bands will tend to flatter younger adults and understate older ones on speed and reasoning items. A child given an adult test, or an adult given a children's test, gets a meaningless number.
Our own test compares you with other people who have taken it, not with your age group: its estimated IQ is that position converted to the mean-100 scale, not an age-normed IQ, because we have not run a representative norming study. If you take the free test, read the result as a rough position within that sample, and see how our test works for what that does and does not mean.