A real IQ test for kids is administered one-to-one by a qualified psychologist, using an age-normed instrument — most often the WISC-V (ages 6–16), the WPPSI-IV (ages 2½–7) or the Stanford-Binet 5. A child's score compares them only to other children of the same age in months, which is why an adult test cannot produce a meaningful number for a child and why free online quizzes are entertainment rather than measurement. This guide covers what the real tests are, when testing is worth doing, what the scores mean, and what to do instead if you are simply curious.
How to check your child's IQ level
There are two routes, and which one fits depends on why you are asking.
Through the school. If the question is about learning support, a suspected difficulty, or eligibility for a gifted program, start by asking the school for an evaluation. In many places an assessment tied to educational need is provided at no cost, carried out by a school psychologist, and comes with the thing that actually matters — a recommendation about what to do next.
Privately. A clinical or educational psychologist can assess independently, usually within a few weeks and for a fee that varies widely by country and practice. This is the faster route, and the one most parents take when a school has a long waiting list or when they want an opinion outside the school system.
What you cannot do is get a defensible number from a website. Scoring a child requires norms for their exact age band — the raw-to-score conversion for a child aged 8 years 2 months is not the conversion for one aged 8 years 11 months — plus a trained examiner watching how the child works, not just what they answer.
The tests professionals actually use
Four instruments cover almost all childhood cognitive assessment in English-speaking countries. The names below are used descriptively: each belongs to its publisher.
| Test | Ages | What it produces |
|---|---|---|
| WISC-V (Wechsler Intelligence Scale for Children, 5th ed.) | 6–16 | Full Scale IQ plus five index scores: Verbal Comprehension, Visual Spatial, Fluid Reasoning, Working Memory, Processing Speed |
| WPPSI-IV (Wechsler Preschool and Primary Scale of Intelligence, 4th ed.) | 2½–7 | Full Scale IQ with a reduced battery designed for very young children and short attention spans |
| Stanford-Binet 5 (SB5) | 2 upwards | Full Scale IQ across five factors, with verbal and nonverbal scores reported separately |
| Nonverbal tests (e.g. NNAT, UNIT) | Varies | Reasoning measured through shapes and patterns, for children with speech differences or who are learning the test language |
The WISC is the one most parents encounter, and the one most often meant by "the IQ test" in a school context. Its five index scores are more useful than the headline number: a child whose Verbal Comprehension sits well above their Processing Speed has a profile that explains far more about classroom life than a single Full Scale IQ ever could.
One thing to be clear about up front: IQ Revealed is not affiliated with Pearson, Riverside Insights or any publisher of the tests described here, does not administer or score them, and does not prepare children for them. We describe them because parents searching for a children's IQ test deserve to know what the real ones are.
At what age can a child be tested?
From about two and a half years old with the right instrument — but "can be" and "is worth" are different questions.
Scores taken in infancy and the toddler years predict later ability poorly. What a two-year-old can be measured on is not really the same construct that an IQ test captures at twelve, and attention, mood and cooperation contribute enormous noise at that age. Stability improves markedly from around age six, once children can sustain a structured session, and by adolescence a well-administered score is reasonably durable.
The practical implication: if there is a pressing reason to test early — a suspected developmental difficulty, an early-entry decision — test early and treat the number as provisional. If the motivation is curiosity, waiting until six or later produces a far more meaningful result, and costs nothing in the meantime.
What is a normal IQ for a child?
Exactly what it is for an adult. The average is 100 by construction, about 68% of children fall between 85 and 115, and about 95% fall between 70 and 130. Children's tests are age-normed, which means the average is pinned to 100 within every age band — so there is no separate children's scale, and no table that converts a child score into an adult one.
| Score | Percentile | Roughly |
|---|---|---|
| 85 | 16th | 1 child in 6 scores at or below |
| 100 | 50th | The exact midpoint |
| 115 | 84th | 1 child in 6 scores at or above |
| 120 | 91st | 1 child in 11 |
| 130 | 98th | 1 child in 44 — the usual gifted threshold |
This is why "is 120 good for a 7-year-old?" has a slightly deflating answer: it is good in the same way a 120 is good at any age, because the adjustment for age has already happened inside the score. A 7-year-old at 120 is at roughly the 91st percentile of 7-year-olds. The full IQ scale and its bands apply unchanged, and you can look up any specific score on the IQ percentile calculator.
Treat any single score as an estimate, not a fact about the child. Measurement error on a well-built test runs roughly ±5 to ±7 points at 95% confidence, so a reported 118 and a reported 124 are not meaningfully different results — how accurate IQ tests are covers where that number comes from.
Are free online IQ tests for kids reliable?
No, and it is worth being direct about why, because a lot of sites offering "free IQ tests for kids ages 4–16" imply otherwise.
Three things are missing from every web quiz. There are no age-band norms — a genuine child score needs a reference sample of children the same age in months, which requires thousands of standardization cases per band. There is no standardized administration — a parent hovering, a second attempt, a hint, or a child clicking through bored all change the result, and a trained examiner exists partly to prevent exactly that. And there is no observation, which is often the most valuable part of a real assessment: how a child approaches a problem they cannot solve tells a psychologist more than whether they solved it.
None of which means online puzzles are worthless. They are decent practice, they can build familiarity with the format before a real assessment, and children often enjoy them. Just do not record the number anywhere or repeat it to the child as their IQ.
What a score is actually used for
A cognitive assessment earns its keep when it answers a question somebody is going to act on. In practice that is usually one of three: whether a child qualifies for a gifted or accelerated program; whether an unexpected gap between ability and achievement points to a learning difficulty worth supporting; or whether a child needs accommodations that a school requires documentation to provide.
Gifted thresholds cluster around IQ 130, the 98th percentile, though this varies enough between districts that the local criteria are worth reading before testing for that purpose. Many programs have also moved away from a single cutoff toward combining scores with teacher ratings and portfolios — a reasonable response to the fact that one number on one morning is a thin basis for a decision about a nine-year-old.
What a score is not for is ranking a child, predicting their life, or settling an argument. It measures a set of reasoning skills at one point in time, and it does so with real error bars. Decisions about a child's education belong to the psychologist who ran the assessment and the people who teach them.
If you are the one who is curious
A large share of people searching for a children's IQ test are parents who are partly curious about themselves — and that is a question we can actually answer.
IQ Revealed's test is built and normed for adults aged 18 and over. We do not offer a children's test, and our terms require account holders to be 18 or older, so please do not use it to test a child — the score would be wrong in a way that looks convincing, which is the worst kind of wrong. For an adult it is a matrix-reasoning self-assessment in the same visual format the supervised tests use: free to take, with membership unlocking the full report, percentile and score breakdown. It is a self-insight assessment rather than a clinical instrument, and our methodology sets out exactly how it is built and scored.
If what you want is context rather than a test, the average IQ and how IQ varies by age are the two places to start.