For most people, a paid online IQ test is not worth it. The price rarely buys better norms, better questions or a more accurate score. More often it buys access to a result you have already earned, plus a certificate. A good free test and a typical paid one share the same limits: no examiner, no controlled conditions and usually no representative norming sample. If you need a score that carries clinical or educational weight, the meaningful upgrade is an assessment by a licensed psychologist, which costs much more and does much more.
How to judge any online IQ test
Price is a poor signal of quality. A better yardstick is what professional testing standards, published jointly by the AERA, APA and NCME, expect of any test: documented norms, evidence of reliability, and a clear statement of intended uses. Translated for online tests, that means five checks:
- It says who you are compared with. A score only means something relative to a reference group. A trustworthy test says whether that group is a representative norming sample or simply its own visitors. If a site reports an "IQ" with no word about norms, assume it has none.
- It reports a range, not a single point. Every test has measurement error. A result shown as "118" with no interval implies a precision the test does not have.
- It shows the full result without a paywall or email gate. If you must pay or hand over an email address to see the score you just earned, what you are buying is access to your own answers, not a better measurement.
- It respects your privacy. Your age can matter for the comparison. Your name, phone number and home address do not.
- It explains what it measures and what it cannot. A good test describes its question types and domains and states plainly that it is not a clinical assessment.
Red flags
- Blurred or locked results. A score teased on screen and hidden behind a payment is a sales funnel.
- Certificate upsells. A printable certificate from an unsupervised online test carries no weight with schools or employers, and high-IQ societies generally accept only supervised testing.
- Inflated scores. If nearly everyone you know scores 125 or higher, the test is flattering its users. On a properly normed test, only about 1 person in 20 scores 125 or above.
- Precise numbers from tiny tests. A 10-question quiz cannot reliably tell 131 from 136. Exact-looking scores from very short tests are decoration.
- Subscription traps. A one-time "report" that quietly renews every month.
- Clinical-sounding claims. Phrases like "clinically validated," "certified IQ" or "accepted worldwide" on an unsupervised test should make you more skeptical, not less.
When paying online can make sense
A fee is not automatically a bad sign. Paying can be reasonable when the money buys something you can check: a longer test, a published technical report describing who was in the norming sample and when, reported reliability and standard error, and a result you can see before deciding whether to buy any extras. Supervised admission tests run by organizations that require them also usually charge a fee, and there the fee pays for a proctor. What a fee should never buy is the score itself, a certificate, or a more flattering number.
Free, paid and clinical tests compared
| Free online test | Paid online test | Clinical assessment | |
|---|---|---|---|
| Who gives it | You, unsupervised | You, unsupervised | Licensed psychologist or trained examiner |
| Comparison group | Rarely representative; good ones say so | Rarely representative; often not disclosed | Representative national norms |
| Typical length | About 5 to 40 minutes | Similar | Hours, including interview, testing and feedback |
| What you get | A percentile or estimate, ideally with a range | Often a single number and a certificate | Full-scale IQ, index scores, confidence intervals and a written report |
| Valid for diagnosis, school or legal decisions | No | No | Yes |
Short vs full IQ tests
Length matters more than price. Each question is a small, noisy sample of your ability, and more questions average out more of the noise. The Spearman–Brown formula, a result from 1910 that is still standard in psychometrics, describes how reliability grows with test length. As an illustration, if a 12-item test had a reliability of .60, a comparable 60-item version would be expected to reach about .88.
Short tests are fine for a quick, rough read. They are not fine for small distinctions, and they are especially weak at the level of individual domains. Our own test shows the trade-off: the 12-question short form (about 5 minutes) has two items per domain and reports its result with a range of ±12 percentile points, while the 60-question full form (about 30 minutes) has ten items per domain and a ±7 range. Those widths are conservative fixed values rather than a measured standard error, but the difference between them reflects a real gap in precision. If you take only one, take the longer one.
What to expect from an online IQ test
A reasonable online test will:
- Take somewhere between 5 and 40 minutes, depending on length.
- Mix question types: pattern matrices, number series, verbal analogies, spatial rotation, memory and speed tasks. Our guide to types of IQ questions has an original example of each.
- Include items of varying difficulty, so expect some you cannot solve. That is by design.
- Either time you or not. Timed tests fold speed into the result; untimed ones score accuracy only.
- Show your result right away, with a clear statement of what it is compared against.
What it will not do is supervise you, verify who is taking it, stop you from using a calculator or search engine, or adjust for a noisy room or a bad night's sleep. Treat the result accordingly. A few sensible steps for getting a fair result are in how to prepare for an IQ test, and the broader question of accuracy is covered in are IQ tests accurate.
When a clinical assessment is worth it
Seek a professional assessment when the score will feed into a decision:
- Evaluation for intellectual disability or a specific learning disability.
- School placement, gifted-program admission or testing accommodations, when the school or program requires a qualified examiner.
- Concerns about cognitive change after a brain injury or illness, or with aging, usually as part of a broader neuropsychological evaluation.
- Legal or disability-benefit settings that require documented testing.
What it involves: a licensed psychologist (or, in a school, a school psychologist) interviews you or your child, gives an individually administered test such as the WAIS-5 for adults or the WISC-V for children, often adds measures of academic achievement, attention or memory, and then writes a report and goes through it with you. The intelligence test itself is commonly an hour or two of one-on-one work. A full psychoeducational or neuropsychological evaluation takes considerably longer and may be spread over more than one session.
Costs vary widely by region, provider and how much testing is included, so ask for a quote and check whether your health insurance covers it; coverage is more likely when there is a medical or diagnostic reason. In the US, if a child is suspected of having a disability that affects learning, the public school system must evaluate the child at no cost to parents under the Individuals with Disabilities Education Act (IDEA). Gifted evaluations follow state and district rules, which differ a great deal.
This is general information, not medical or educational advice. A psychologist can tell you whether testing would actually answer your question.
Which option to choose
If you are curious how you do on reasoning puzzles, a transparent free test is enough, ideally a longer one that shows a range and tells you who you are compared with. You can take the free test here; it needs no account, email or payment, and its estimated IQ is labelled as exactly that: your position among its own respondents, converted to the familiar scale and shown with a range, not a clinical score. If you need a score for a diagnosis, a school or a legal process, skip online tests entirely and go to a professional. There is rarely a good reason to pay for the option in between.