Which IQ test to take is a question about what the score is for

0 The short answer

Which IQ test you should take depends on what the score is for: for curiosity, a free screener or the ACIS free trial is enough; for a profile of your strengths across several abilities, an adult-normed, multi-domain online battery is the right class, and ACIS offers one in three forms (Quick at 25 dollars in about 45 minutes, Optimized at 35 dollars in about 110 minutes, Full Scale at 59 dollars in about 175 minutes); and for Mensa, a giftedness question, a school decision or a clinical question, only the party that will accept the result can give you a score that counts. Mensa's own admission test lists at 60 dollars for a Local Group session and 99 dollars for private testing, and one university clinic lists a Wechsler gifted assessment at 100 to 400 dollars. All prices on this page were read on November 4, 2026 and can change. The decisive limit is that an unsupervised online score, ACIS included, is a personal estimate and not a credential.

1 Why Does the Right IQ Test Depend on What the Score Is For?

An IQ test is not good or bad in general; it is valid or not valid for a stated use, which is why the same test can be the right choice for one reader and the wrong one for another. The Standards for Educational and Psychological Testing, the joint document of the American Educational Research Association, the American Psychological Association and the National Council on Measurement in Education, opens its chapter on validity with Standard 1.0: each intended score interpretation for a specified use should be clearly articulated, and appropriate validity evidence should be provided for each. The comment on Standard 1.1 adds that statements about validity should refer to particular interpretations and uses. Read plainly, that means a score has no meaning until you say what you will do with it.

That is why a ranking of products cannot answer this question. A test can be an excellent estimate of where you stand among adults and still be useless as a document for a school, a clinic or a society, because those parties set their own rules about what they accept. The reverse also holds: a supervised society exam can admit you to a society and still tell you nothing about your profile, because, as the American Mensa testing page states, it is given for admission and "not to quantify intelligence." Three questions sort every option on this page:

1. Who is the score compared with? A score is a position in a reference group. Standard 5.8 says norms, if used, should refer to clearly described populations, including the people with whom users ordinarily wish to compare examinees. The page on how IQ scores are normed explains what that means in practice.
2. How many tasks stand behind it? The Standards state in their chapter on reliability that if an assessment is shortened, for example by decreasing the number of items or tasks, reliability is likely to decrease, and that lengthening it with comparable tasks is likely to increase it.
3. Will the party that asks for the number accept it? That is a rule of the institution, not a property of the test, and Standard 1.4 puts the burden on the user: if a score is interpreted for a use that has not been validated, the user has to justify it.

Four neighboring pages own questions that this page does not repeat. The comparison of products on quality grounds is on the ranking of the best online IQ tests. The taxonomy of instruments is on the page about the types of IQ tests. How to read a score you already have is on What Is My IQ, and the list of physical and online places is on where to take an IQ test. This page does one thing: it decides, by goal, which class of test to take, and what that route cannot do.

2 How Do the Real Options Compare?

The options fall into five classes, and the useful difference between them is not quality in the abstract but who compares you with whom, whether anyone supervises, and what you can do with the result. The table lists what each class measures, who the score is compared with, whether it is supervised, how long it takes, what it costs and what you receive. Facts about other publishers come from their own pages, opened on November 4, 2026.

Four facts in the table deserve a comment. First, the Pearson completion time of 45 minutes belongs to an instrument that the same page describes as individually administered, so the minutes are not comparable with an unsupervised session. Second, the clinic fees are the fees of two university clinics we read, not the price of the test, and a private practice may charge differently; the page on how much an IQ test costs has the wider price breakdown. Third, the Mensa row is the only one whose purpose is admission, and the page says outright that it cannot give you a score. Fourth, the free row is not a ranking of products: it covers a range from honest entertainment to research instruments, and the research instrument we opened, ICAR, states on its own project page that no norms were collected for it.

The comparison between professional and online testing has its own page, professional versus online IQ tests. Here the point is narrower: which class fits which goal, which the next section sets out.

3 Which Test Fits Each Goal?

Each goal maps to one class of test, and the mismatch to avoid is paying for a route that cannot do what you need, such as a society exam when you want a profile or an online battery when you need a document. The table below decides by goal. The columns say what to take, how long it takes, what it costs (read on November 4, 2026 unless noted) and what that route cannot do.

Read the table from the left column. If the goal is one of the last four rows, the first question is not which test but who decides, and the answer is on that party's own page. If the goal is one of the first four rows, the choice is yours, and the next two sections explain how to make it.

One consequence is easy to miss. The first four rows are all served by the same kind of instrument, an adult-normed online battery, in different amounts. A short form is a cheaper and faster answer to a narrower question, and a long form is a slower and costlier answer to a wider one. The page on a quick IQ test covers what a short battery costs you in precision, and the page on a comprehensive IQ test covers what complete really means. The rest of this page assumes you have chosen the goal and now need the amount.

4 What Should You Take If You Are Only Curious?

For curiosity, a free tool is a reasonable first step provided you read the number as a position among the visitors of that tool and not among adults in general. There is nothing wrong with wanting a rough sense of where you stand, and the cheapest honest way to get one is to spend nothing. Two things decide whether the free route is useful.

The first is the reference group. A free site that states whom it compares you with has told you what the number means. One that does not has not. The Standards, in Standard 5.8, ask that norms refer to clearly described populations, and a score computed against everyone who took a quiz is a rank in a crowd that nobody sampled on purpose. The page on free IQ tests covers the free options and what each gives you.

The second is whether the instrument was ever meant to give an individual a number. The International Cognitive Ability Resource, ICAR, is a good example of a free instrument that was not. Its project page, hosted by the Department of Statistics at TU Dortmund University and read on November 4, 2026, describes it as a large collection of assessment instruments for cognitive abilities and states that it is "intended for academic use exclusively." Asked whether there are norms, its FAQ answers no: there was no attempt to use existing data or to collect data specifically for that purpose. Asked whether you can infer an individual IQ from it, the FAQ answers that you should not, and it says that the ICAR 60 cannot replace a proper cognitive ability assessment under controlled conditions. That is a clear label, and it is a reminder that free and public do not mean normed.

If you want a free look that is also an adult-normed instrument, ACIS has a free trial that needs no card. The site allows one free trial per person and refuses a second one from the same email address or the same browser. Finished trial subtests are kept if you later unlock a form that includes them, so the trial is a way to see the task formats and the report before you decide, not a throwaway. The page on the IQ test for adults describes where ACIS fits for an adult who wants more than a number.

What the free route cannot do deserves a plain sentence. It cannot give you a score that anyone else must accept, and for most free tools it cannot tell you how precise the number is. If the number will change a decision, move down the table to a route that reports a band around the score. The sibling page on a legit IQ test covers what separates a test worth trusting from one that only looks like one.

5 What Should You Take for a Profile of Your Strengths and Weaknesses?

For a profile of strengths and weaknesses, take a battery that samples several broad abilities, because one number cannot show a profile and a single subtest cannot show an ability. The framework most professional batteries use to decide what a battery should cover is the Cattell-Horn-Carroll theory. McGrew (2009) describes how the Cattell-Horn Gf-Gc and Carroll Three-Stratum models emerged as the consensus psychometric models of the structure of human intelligence, and how the term CHC theory names a synthesis of the two. The Woodcock-Johnson V, a major individually administered battery, states in its publisher technical abstract that it measures cognitive abilities as defined by the contemporary Cattell-Horn-Carroll theories. The CHC model page explains the broad abilities one by one.

In CHC terms the six ACIS indices are Gc (verbal comprehension, the VCI), Gf (fluid reasoning, the FRI), Gq (quantitative reasoning, the QRI), Gv (visual spatial processing, the VSI), Gwm (working memory, the WMI) and Gs (processing speed, the PSI). The page on what IQ measures explains why these are separate abilities and how they relate to general ability. What differs between the three ACIS forms is how many of the six they cover:

The domain columns follow from the subtest lists of each form. Seven subtests appear only in the Full Scale form: Information, Arithmetic, Mathematical Achievement, Visual Number Series, Complex Relations, Visual Puzzles and Visual Sequence. That has a practical meaning. If quantitative reasoning matters to you, only the Full Scale form or a Custom Test with Arithmetic and Mathematical Achievement will report it, and if the visual spatial index matters, Optimized covers it through Spatial Comprehension and Layer Rotation, while Full Scale adds Visual Puzzles.

A few subtests show how the formats differ without reproducing any content. Matrix Reasoning asks you to solve visual pattern problems, which sample abstraction, logic and nonverbal reasoning. Vocabulary measures word knowledge through definition questions. Digit Span uses forward and backward digit recall, and, like Alphanumeric Sequencing and Arithmetic, it presents its items as audio and needs working sound. The speed subtests, Coding and Symbol Search, ask for fast, accurate matching and scanning of symbols. Layer Rotation, as the technical manual states, carries no words at all.

When you read a profile, remember that a gap between two indices is not a diagnosis. The page on index score differences covers how common gaps are, and the page on the General Ability Index covers one of the composites the manual lists. The technical manual lists further composites built from predefined subtest sets, available when their subtests are available. It describes the CPI, for example, as combining "working memory and processing speed as cognitive proficiency." The technical manual is the place for the documentation.

Custom Test is the route for readers who want one domain and nothing else. It costs 10 dollars per subtest, and the site states that any domain you complete returns its full index score and every subtest returns its scaled score. The cost of a domain is 10 dollars times its subtests (our arithmetic): 20 dollars for quantitative reasoning or processing speed, 30 for visual spatial or working memory, and 50 for verbal comprehension or fluid reasoning. From 6 subtests the Full Scale form costs less, since six Custom subtests cost 60 dollars and the Full Scale form costs 59 (our arithmetic), and only the Full Scale battery returns a Full Scale IQ. The example report shows what the output looks like before you pay.

6 What Should You Take If the Question Is Giftedness?

For a giftedness question, the test that counts is an individually administered battery from a psychologist, and an online battery is the cheaper first look at where you stand, not a substitute. Giftedness in adults has no single official cutoff, which is why the word needs care before you choose a route. A common reference point is an IQ of 130, which sits two standard deviations above the mean of 100 on a scale with a standard deviation of 15, about the 98th percentile (our arithmetic from the normal curve). American Mensa uses a cutoff at a similar level and states on its testing page that a qualifying score indicates you have tested at or above 98 percent of the general population. The gifted test for adults page works through that question.

A psychologist offers three things that no online route does. The first is supervision: the examiner sees how you work and records conditions. The second is the document, which a school, a clinic or a program may require. The third is judgment: the Standards in Standard 10.1 say those who use psychological tests should confine their activities to their areas of competence, shown through education, training, experience and credentials, and Standard 10.2 asks that they be familiar with the evidence of validity and reliability for the intended uses. The Pearson WAIS-5 page lists a qualification level of C, and the page on who can administer an IQ test explains what the levels mean. The WAIS test online page covers the official options and why a Wechsler instrument cannot be taken at home.

What does it cost? We opened two university clinics on November 4, 2026. The University of South Florida Psychological Services Center lists a "Gifted Assessment" using the Wechsler Intelligence Scale for Children or the Wechsler Adult Intelligence Scale, with an evaluation range of 100 to 400 dollars, and a deposit of half the quoted fee at the first meeting; it says fees for assessment-only services depend on income. The page does not say who administers it or what the deliverable is. The Emory University Psychological Center lists a comprehensive cognitive assessment at 900 to 1,800 dollars on a sliding scale based on income, conducted over several sessions that can each last 2 to 4 hours, in a training clinic staffed by graduate students in clinical psychology who work under licensed clinical psychologists. Those are two data points, not a market average, and a private practice may charge more or less.

Where does an online battery fit? If you only want a first look, an adult-normed profile tells you whether the question is worth the clinic fee, and it costs 35 or 59 dollars for the multi-domain forms. It cannot confirm giftedness for anyone else, and the ACIS site says it is not for accommodation determinations or admission to high IQ societies. For readers who ask about scores at the top of the range, the page on high range IQ tests covers why the top of any scale needs separate care, and the sibling page on the intelligence test explains what the word covers. If the question behind your question is about a diagnosis of any kind, the right person to ask is a psychologist, and no score on this page should replace that conversation.

7 What Should You Take If You Want to Join Mensa?

To join Mensa, take the Mensa admission test itself, because no online score from any other provider is a substitute for it and this page does not claim otherwise. The American Mensa testing page, read on November 4, 2026, lists two options at regular prices: Local Group Testing at 60 dollars, with sessions proctored by Local Group volunteers, and Private Testing at 99 dollars, at more than 400 approved testing facilities, where the exam is administered electronically. The page notes that promotional discounts appear in the shopping cart at checkout, and we print only the regular prices. It says the test takes between one and two hours.

Two sentences from that page set the expectations. The first is the cutoff: a qualifying score indicates that you have tested at or above 98 percent of the general population. The second is the purpose: the admission test is given for admission into Mensa and not to quantify intelligence, and Mensa states that it cannot provide a detailed report with scores, percentile ranks or an IQ score. Whether you pass is a yes or no. That is the opposite of what a profile needs, and it is why a reader who wants to know where they stand has to look somewhere else. The page on the Mensa IQ test covers what score the cutoff corresponds to and how to estimate yours.

Where does an online battery help? It is not a Mensa test, it does not predict a result, and nothing on this page suggests that Mensa accepts it. What it can give you is an adult-normed picture of where you stand before you pay a fee, including a band around the number. A reader whose Full Scale IQ and band sit far below the cutoff has learned something cheaply, and a reader close to it has learned that the decision will be close. For any score taken elsewhere, including a professional report, ask Mensa directly what it accepts; this page makes no claim about it. The ACIS limit is stated once, in the last section, in the words of the site.

8 What Should You Take for a Clinical, School, Accommodation or Workplace Question?

For a clinical, school, accommodation or workplace question, the right test is whichever one the deciding party accepts, which in practice means a licensed psychologist or the employer's own selection process, and not a consumer IQ score. The Standards are explicit that the decision, not the score, is the unit. In educational settings, Standard 12.10 says a decision or characterization with major impact on a student should consider not just scores from a single test but other relevant information. The introduction to the same chapter adds that test scores alone should never be used as the sole basis for including a student in, or excluding a student from, special education programming.

Employment is a separate case. Standard 11.3 says that when test content is a primary source of validity evidence for employment decisions, a close link between the content and the job requirements should be demonstrated. A general IQ score from a consumer site comes with no such link, and a person who takes one should not expect an employer to treat it as a selection result.

The APA guidance points the same way. The APA Guidelines for Psychological Assessment and Evaluation, approved by the APA Council of Representatives in March 2020, ask psychologists in Guideline 6 to select tools that demonstrate sufficient validity evidence for their uses and sufficient score reliability. Elsewhere their rationale says individual performance on psychological tests is only one piece of an assessment, set in a context of presenting concerns, history and information from multiple sources. A consumer test is, by its design, one piece with no context, and the guidelines are written for psychologists, though a reader can borrow the same checklist.

If a clinical or school question is behind your search, there are three practical steps. Ask the party that will use the result, in writing, which instruments and which examiners it accepts. Ask a licensed psychologist whether a Wechsler or Woodcock-Johnson battery is appropriate for your question, and what the fee covers. And treat any online score, ACIS included, as one input that may help you decide whether to ask, not as an answer. The Standards, in Standard 10.5, ask that tests suit the characteristics and background of the test taker, including linguistic characteristics, and the comment adds that when no normative or validity studies exist for a relevant population, interpretations should be qualified and presented as hypotheses rather than conclusions. That is a standard for professionals, and it is also good advice for anyone reading their own result.

9 How Long Should the Test Be, and Does Coverage Matter?

A longer test gives a more precise score and a wider profile, so the right length is the shortest one that answers your question with a band narrow enough to act on. The Standards state in their chapter on reliability that if an assessment is shortened, by decreasing the number of items or tasks for example, reliability is likely to decrease, and that lengthening it with comparable tasks is likely to increase it. The same chapter calls lengthening an effective and commonly used way to improve reliability. That is general psychometrics and not an ACIS statistic, and the page on reliability and validity explains the terms.

Length is only part of it. A test can be long and narrow, or short and broad. A Full Scale IQ from a battery that samples three domains is a summary of three kinds of tasks, and one from six domains is a summary of six. The Pearson page for the WAIS-5 shows how a professional battery handles this: seven subtests build the Full Scale IQ, ten subtests are the primary index subtests, and the publisher lists administration times of 45 and 60 minutes for those two sets, for an examiner working with one person. Online administration is different, since nobody moves the session along.

For the ACIS forms the trade is explicit. Quick is the shortest, at about 45 minutes, and is built for a rough number with a band. Optimized takes about 110 minutes and adds visual spatial and processing speed domains, giving five domains. Full Scale takes about 175 minutes and covers all six. The forms allow breaks, save progress between subtests and give 30 days to finish, so a longer form does not need to be taken in one sitting. For readers who weigh the long route, the sibling page on the long IQ test is the place to compare. A practical rule follows from all this. If the answer you want is a number, start with the shortest form that reports a band. If the answer you want is a profile, pay for the domains you want to see, and no more.

10 How Do You Read the Score You Get?

Read an IQ score as a position in a described reference group with a band around it, never as a fixed property of you, and read a single subtest as the least precise number in the report. ACIS reports indices and the Full Scale IQ on the standard scale, with a mean of 100 and a standard deviation of 15, and subtests as scaled scores from 1 to 19, with a mean of 10 and a standard deviation of 3, together with percentiles and a 95 percent confidence interval. The page on IQ score versus percentile explains why the two scales cannot be averaged, and the percentile and rarity calculators let you convert a score once you have one. A band is part of the result, and a number printed without one hides how much it could move.

A single subtest is less precise than a battery for the reason the Standards give: fewer tasks, less reliability. An index built from two or three subtests sits between the two. A profile is therefore a set of estimates with different bands, and a small difference between two indices tends to mean less than a large one. That is a general principle and not a statement about any ACIS index.

Three more things change what a score means. The first is the age of the norms. A meta-analysis by Trahan, Stuebing, Fletcher and Hiscock (2014) covering 285 studies and 14,031 participants found a meta-analytic mean Flynn effect of 2.31 standard score points per decade, and 2.93 points per decade across 53 comparisons involving modern Stanford-Binet and Wechsler tests. Because scores on a given test rise as its norms age, the same performance reads higher against older norms, and a recently normed test is the safer choice. The second is repetition. Scharfen, Peters and Holling (2018) pooled 174 samples from 122 studies with an overall sample of 153,185 and found significant retest effects and no further score gains after the third test administration, so a second attempt at the same kind of test is not a clean second measurement. The third is conditions: an unsupervised session is the subject of the next section.

11 What Conditions Does an Unsupervised Session Need?

An unsupervised session needs working sound, a stable device, a quiet room and a plan for breaks, because nobody else is there to standardize the conditions and the score carries every shortfall. The Standards, in the comment to Standard 3.4, note that computerized testing adds concerns for standardization: examinees need access to technology so that aspects of the technology itself do not influence scores, and examinees working on older, slower equipment may be unfairly disadvantaged. They also say that unproctored administrations, where standardization cannot be ensured, could give some test takers an advantage over others, and that in these situations results should be interpreted with caution. That is the honest limit of every online route, and the reader is the only person who can reduce it.

The practical list is short. Use a laptop or desktop with a reliable connection. Test the sound before you start, because Digit Span, Alphanumeric Sequencing and Arithmetic present their items as audio and need it to work. Choose a time when you are rested and nobody will interrupt you, and silence notifications. The forms allow breaks and save progress between subtests, so take them between subtests and not in the middle of one. If a task has a time limit, the subtest page and the technical manual say what the rules are, and where they do not state a detail, this page does not either.

Two conditions are not under your control. The test is in English only, and the technical manual says verbal and knowledge tasks are affected by English proficiency and educational exposure, so a reader who is learning English should read the Vocabulary or Information results with that in mind. And the session is unsupervised, so nobody confirms who took it or how.

12 Where ACIS Sits, and What a Buyer Should Do

ACIS is a route for an adult who wants a multi-domain, adult-normed profile online at a low one time price, and it is not a route to a document, a diagnosis or a society, because it is an online, unsupervised, English only test that is not a clinical instrument. We sell ACIS, so treat this section as a disclosure and check it against the rest of the page and against the documentation.

What ACIS offers is stated plainly. There are three one time forms, Quick at 25 dollars, Optimized at 35 dollars and Full Scale at 59 dollars, plus a Custom Test at 10 dollars per subtest, with no subscription. Every form reports a Full Scale IQ with its 95 percent confidence interval, the indices and percentiles are on the standard scale, and adult norms cover ages 16 to 90. Breaks are allowed, progress is saved between subtests and a buyer has 30 days to finish. The free trial needs no card and is limited to one per person. The refund condition, quoted literally, is "Full refund within 5 days if the report does not deliver the features described at checkout or a technical issue prevents you from accessing your results." That is a conditional refund and not an unconditional money back guarantee. Prices were read on November 4, 2026 and can change again.

What ACIS does not offer is also stated plainly. The research page of the site says ACIS "is not intended for clinical diagnosis, hiring decisions, accommodation determinations or admission to high IQ societies." The test is taken online and unsupervised, so nobody confirms who took it or under what conditions. It is available in English only, and the technical manual says verbal and knowledge tasks are affected by English proficiency and educational exposure. The full technical documentation is where the evidence lives, and we point there instead of quoting statistics on this page.

The documents that govern how a score should be read agree with these limits. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), which we read in the open access edition, say in Standard 1.0 that each intended score interpretation for a specified use should be clearly articulated and supported by validity evidence. In the comment on Standard 3.4 they say that unproctored administrations, where standardization cannot be ensured, could give some test takers an advantage over others and that results in those situations should be interpreted with caution. Standard 5.8 asks that norms refer to clearly described populations, and Standard 9.4 makes the user responsible for the rationale and the evidence when a test is used for a purpose with little or no validity evidence. The APA Guidelines for Psychological Assessment and Evaluation ask psychologists in Guideline 6 to select tools with sufficient validity evidence for their uses and sufficient score reliability, and elsewhere they ask them to consider characteristics such as the standardization or normative group when judging whether a test applies. Those guidelines are written for psychologists, and a buyer can borrow the same checklist: for what purpose, compared with whom, and how precise.

For a concrete next step, decide first what the score is for. If it is curiosity, start the free trial, which needs no card, and open an example report before you spend anything. If it is a personal estimate with a stated band, choose Quick at 25 dollars. If it is a profile, choose Optimized at 35 dollars for five domains or Full Scale at 59 dollars for all six, and take it in a quiet room with working sound. If it is one ability, use a Custom Test. If it is Mensa, book the Mensa test. If it is a giftedness, clinical, school, accommodation or hiring question, ask the party that will use the result what it accepts, and pay for that.

13 Sources Behind This Page

Every figure on this page that is not our own arithmetic is traceable to one of the following, and each is linked at the point where it is used. Web pages were opened on November 4, 2026, and their prices and statements can change. Totals, the 98th percentile for an IQ of 130, and the cost of a Custom Test domain are our arithmetic on published numbers and are labeled as such. Competing online IQ products are not used as sources or linked.

• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing. American Educational Research Association, 2014 (open access edition), read on November 4, 2026.
• American Psychological Association, APA Task Force on Psychological Assessment and Evaluation Guidelines. APA Guidelines for Psychological Assessment and Evaluation. American Psychological Association, approved March 2020, read on November 4, 2026.
• American Mensa. Testing options and fees, page read on November 4, 2026.
• Pearson. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5) product page, page read on November 4, 2026.
• LaForte E M, Dailey D, McGrew K S. Woodcock-Johnson V Technical Abstract. Riverside Assessments, 2025, read on November 4, 2026.
• University of South Florida, Psychological Services Center. Clinic policies: assessment fees, page read on November 4, 2026.
• Emory University Psychological Center. Frequently asked questions: cost of a comprehensive assessment and training clinic staffing, page read on November 4, 2026.
• TU Dortmund University, Department of Statistics. International Cognitive Ability Resource (ICAR) project page and FAQ, page read on November 4, 2026.
• McGrew K. CHC theory and the human cognitive abilities project: Standing on the shoulders of the giants of psychometric intelligence research. Intelligence, 2009, volume 37, issue 1, pages 1 to 10.
• Trahan L, Stuebing K, Fletcher J, Hiscock M. The Flynn effect: A meta-analysis. Psychological Bulletin, 2014, volume 140, issue 5, pages 1332 to 1360.
• Scharfen J, Peters J, Holling H. Retest effects in cognitive ability tests: A meta-analysis. Intelligence, 2018, volume 67, pages 44 to 66.

14 Frequently Asked Questions