Most online IQ tests are not accurate enough to give you a real IQ score. They usually lack a representative norm sample, run unsupervised, can be retaken, and some inflate results to sell certificates. A few are reasonable as rough indicators, but only an individually administered, standardized test (such as the WAIS or WISC) given by a psychologist yields a valid IQ — with a margin of error of about ±5–6 points.
Why online IQ tests are usually inaccurate
- No real norm sample. An IQ score is a rank against a representative group of people your age. Many websites never test such a group; others norm on their own self-selected visitors, who differ from the general population.
- Too few, too similar items. Clinical tests combine 10 or more subtests (verbal, visual-spatial, reasoning, memory, speed). Many online tests use 20–40 matrix puzzles only, which measures a narrow slice of ability.
- No supervision. You can pause, look things up, use a calculator, ask a friend or retake the test. Each retake adds practice effects.
- Flattering scoring. Some commercial tests report high numbers to encourage sharing and to sell detailed reports or certificates.
- Ceiling and floor problems. With few very hard items, a test cannot meaningfully distinguish 140 from 160; claims of very high scores are especially unreliable.
- No confidence interval. A serious report gives a range, not just a single number.
Are any online IQ tests accurate?
Some are better than others. Research-grade item banks such as the International Cognitive Ability Resource (ICAR; Condon & Revelle, 2014) are public-domain, have published psychometric properties and are used in large studies — but they are designed for research on groups, not for diagnosing individuals. Some commercial tests report validation against established tests; check whether they publish their norm sample, reliability and correlation with a clinical test, and whether these were peer-reviewed.
How to tell a better online IQ test from a worse one
| Better signs | Warning signs |
|---|---|
| Describes its norm sample (size, ages, countries) | No information about how scores were derived |
| Reports reliability and a confidence interval | A single precise number with no range |
| Several item types and a time limit | Only 10–20 items, untimed |
| Free result, no pressure to pay | Score hidden behind a payment or “certificate” |
| Clear disclaimer that it is not a clinical test | Claims to be “official” or “Mensa-certified” |
| Typical results around 100 for typical users | Most users get 120–140 |
What about our IQ test?
Our free 120-question IQ test includes matrix puzzles, number series, verbal analogies, and maths and logic items with a 60-minute timer and scores by section. We are open about its limits: all items were written for this site, the result is a rough estimate for fun and brain training, and untimed play caps the estimate at 130 because the scoring assumes timed conditions. Use it to practise and to see which item types you find easy or hard — not as a clinical IQ.
Mensa and other high-IQ societies
Mensa requires a score in the top 2% on an approved, supervised test — about 130 on an SD-15 scale. Online results are not accepted for admission. National Mensa groups run supervised entrance tests or accept prior qualifying scores from recognised tests.
How to get an accurate IQ score
- Ask a licensed or registered (clinical, educational or neuro-) psychologist for a full assessment with a standardized test such as the WAIS-5/WAIS-IV, WISC-V or Stanford-Binet 5.
- Expect 1–2 hours of testing and a written report with index scores and a 95% confidence interval.
- Schools, universities (disability or learning-support services) and health services can often refer or provide assessment when there is a clinical or educational reason.
- Sleep well and take the test as you are — preparation or practice on similar items can inflate scores and reduce their usefulness.
Sources
- Condon, D. M., & Revelle, W. (2014). The International Cognitive Ability Resource (ICAR). Intelligence, 43, 52–64.
- Wechsler, D. (2008). WAIS-IV Technical and Interpretive Manual. Pearson. (Descriptive classifications; FSIQ range 40–160; normative sample N = 2,200, ages 16–90.)
- Pearson (2024). WAIS-5 sample score report (shows FSIQ 111, 95% CI 106–116, SEM 3.00).
- Canivez, G. L. (2010). Review of the WAIS-IV. Mental Measurements Yearbook (FSIQ internal consistency .97–.98).
- Mensa International: admission requires a score in the top 2% on an approved, supervised intelligence test.
- Neisser, U., et al. (1996). Intelligence: Knowns and unknowns. American Psychologist, 51(2), 77–101.
Frequently asked questions
Are online IQ tests accurate?
Usually not. Most lack a representative norm sample and supervision. They can be fun or rough indicators, but they do not give a valid IQ score.
Why did I score higher on an online IQ test?
Common reasons include easy items, flattering scoring, no time limit, retakes and norms based on self-selected visitors.
Is there a free accurate IQ test online?
Research tools such as ICAR have documented properties, but they are designed for group research. A valid individual IQ requires a standardized test given by a psychologist.
Does Mensa accept online IQ tests?
No. Mensa requires a top-2% score on an approved, supervised test.
How accurate is a professional IQ test?
Very reliable: full-scale reliability around .97–.98 and an SEM of about 3 points, so the 95% confidence interval is roughly ±6 points.