Every IQ Test Type Compared: Mensa, Raven's, Wechsler, and More

The proliferation of IQ tests is not the result of one test being wrong and the others trying to fix it. It is the result of different pioneers, working in different eras, asking different questions about intelligence — and often building tests for different populations.

Alfred Binet and Théodore Simon built the first practical intelligence test in France in 1905 to identify schoolchildren who needed extra support. Lewis Terman brought that test to Stanford in 1916, rescaled it for American children, and gave us the Stanford-Binet. David Wechsler, working in a New York hospital, thought Binet's test was too child-focused and too verbal, so in 1955 he published the Wechsler Adult Intelligence Scale with a mix of verbal and performance subtests designed specifically for adults. John C. Raven, asking whether a language-free test of abstract reasoning could capture the "g-factor" of general intelligence, published his Progressive Matrices in 1936. Raymond Cattell, worried that verbal tests were culturally biased, built the Culture Fair Intelligence Test in the 1940s using only visual puzzles.

Each of those tests solved a real problem, and each is still in use because the problem it solved has not gone away. That is why a psychologist assessing a learning disability will reach for a Wechsler, a researcher comparing cognitive scores across languages will reach for a Raven's or a Cattell, and a Mensa applicant might encounter any of the above depending on which chapter is administering the test.

Before we get to the side-by-side table, here is a one-paragraph orientation to each of the major tests. These are the names you are most likely to run into when you start reading about IQ assessment.

Wechsler Adult Intelligence Scale (WAIS). First published in 1955 by David Wechsler and now in its fourth revision, the WAIS is the de facto standard for adult clinical IQ assessment in the English-speaking world. It produces a full-scale IQ (mean 100, standard deviation 15) plus four index scores covering verbal comprehension, perceptual reasoning, working memory, and processing speed. Administration is one-on-one with a trained examiner and typically takes 60 to 90 minutes. For a deeper tour of what is actually inside the test, see our dedicated Wechsler guide.

Stanford-Binet Intelligence Scales. The direct descendant of the original 1905 Binet-Simon test, adapted by Lewis Terman at Stanford in 1916 and revised repeatedly since. The current fifth edition (SB5) was published in 2003 and covers ages 2 through adult, which historically made it the go-to test for children before the Wechsler children's scales matured. It uses a mean of 100 and a standard deviation of 15 in its modern form, though older editions used a standard deviation of 16 — a quirk that still shows up in legacy score reports.

Raven's Progressive Matrices. John C. Raven's 1936 nonverbal test asks you to complete visual patterns by picking the missing piece of a grid. It comes in three versions — Standard Progressive Matrices (SPM) for the general population, Coloured Progressive Matrices (CPM) for children and older adults, and Advanced Progressive Matrices (APM) for high-ability testing. Because it uses no words and no culturally specific imagery, Raven's is often described as quasi-culture-fair, and it correlates unusually well with the statistical g-factor. We go deep on how the matrices work in our Raven's Progressive Matrices guide.

Cattell Culture Fair Intelligence Test (CFIT). Raymond Cattell's 1940s contribution to the culture-fairness problem. CFIT uses only visual puzzles — series completion, classifications, matrices, and conditions — and comes in three scales for different age and ability groups. One quirk worth remembering: Cattell's test historically uses a standard deviation of 24, not 15, so a "Cattell 148" and a "Wechsler 130" are both at roughly the 98th percentile even though the numbers look very different. See our breakdown in culture-fair IQ tests explained.

Mensa-administered and -accepted tests. Mensa is not itself a test — it is a high-IQ society whose membership requires a score at or above the 98th percentile on an approved intelligence test. Mensa administers its own supervised group tests and also accepts prior scores from a long list of approved assessments, including the Wechsler, the Stanford-Binet, the Cattell, and certain Raven's forms. Because the qualifying threshold is defined as a percentile rather than a raw number, the cutoff looks different on each test. Our Mensa qualification guide walks through the specifics.

Short online IQ tests. Not a single test but a category: the 5- to 15-minute web assessments that anyone can take from a browser. Psychometrically they range from "reasonable casual estimate" to "barely better than guessing," depending on how the items were selected and how — or whether — they were normed against a reference population. They are the right tool for curiosity and the wrong tool for any decision with real stakes. For an honest look at what short tests can and cannot tell you, read our deep dive on quick IQ test accuracy.

This is the part of the article people usually screenshot. Every row is a test type, every column is a dimension worth comparing. Note that "cost" is a rough range for a single administration in US dollars and varies dramatically by country and provider; "length" is the typical testing session, not the paperwork around it.

Two things jump out of that table. First, the price-to-rigor gradient is steep: you can take a free online test in five minutes or pay hundreds of dollars for a clinical Wechsler, and the gap between those two experiences is enormous. Second, the standard deviation column is not decoration — a score of 140 on a Cattell (SD 24) is not the same percentile as a 140 on a Wechsler (SD 15), even though they use the same mean of 100.

The easiest way to pick an IQ test is to work backward from why you want one. Here is a decision framework based on the most common reasons people start searching.

• "I'm just curious and want a rough number." A short, well-designed online test is the right tool. It takes five to ten minutes and gives you a ballpark estimate. Try our quick IQ test.
• "I want to know what a real Wechsler-style test feels like." You want a longer, mixed-content experience that mirrors the classical structure of a clinical test. Our classical IQ test is built around that model.
• "I'm thinking about applying to Mensa." You need practice on the kinds of items the Mensa supervised tests favor — abstract reasoning, pattern recognition, and number series under a clock. Our Mensa-style test is tuned for that goal.
• "I want nonverbal pattern testing." Raven's Progressive Matrices are the gold standard here. We offer a standard Raven's-style matrix test for the untimed experience, and a timed Raven's-style test if you want to practice under time pressure.
• "English isn't my first language and I want a language-neutral test." A visual, culture-fair assessment is the right choice. Our culture-fair IQ test uses only visual items.
• "I need a clinical diagnosis." See a licensed psychologist. No online test — ours or anyone else's — substitutes for a supervised clinical assessment when the result has to stand up in a medical, legal, or educational setting.

Almost every modern IQ test reports a "deviation IQ" rather than a raw score. That means your number is not how many questions you got right — it is how far your performance sits from the average of a reference population, expressed in standard deviations and then rescaled so the average comes out to 100.

The trick is that the rescaling uses different step sizes on different tests. The Wechsler and modern Stanford-Binet use a standard deviation of 15, which is the contemporary convention. Older editions of the Stanford-Binet used 16. The Cattell Culture Fair Test uses 24. That is why a Cattell score of 148 and a Wechsler score of 130 can both describe someone at the 98th percentile — they are the same distance above the mean, just measured in different units.

If this is the part of IQ testing that interests you most, our piece on how IQ scoring works and the bell curve walks through the math step by step.

No IQ test is truly culture-free. Sitting for a test is itself a cultural activity, and even a purely visual item can be easier for people who have grown up with grids, graphs, and printed puzzles. But tests do land at different points along a spectrum of cultural loading.

• Most culturally loaded: heavily verbal tests that depend on vocabulary, idiom, and general knowledge. These penalize non-native speakers and readers from different educational traditions.
• Middle of the spectrum: the Wechsler and Stanford-Binet, which mix verbal and nonverbal subtests and whose full-scale IQ draws on both.
• Least culturally loaded: Raven's Progressive Matrices and the Cattell Culture Fair Test, both of which use only abstract visual items and avoid language entirely.

Choosing a more culture-fair test does not make the score "objectively better" — it makes it more comparable across language backgrounds. If you want the longer argument about what culture-fair really means and where the concept came from, our culture-fair IQ tests explained piece covers it in depth.

Take a Wechsler on Monday and a Raven's on Tuesday and you will almost certainly get two different numbers. That is expected, not a sign that one of the tests is broken. Different instruments measure overlapping but not identical cognitive constructs. The Wechsler weights verbal comprehension and working memory heavily; Raven's focuses almost entirely on abstract visual reasoning; the Cattell leans hard on fluid intelligence. Someone with strong vocabulary and average pattern recognition will tend to score higher on the Wechsler; someone with weaker language skills and sharp abstract reasoning will tend to do better on Raven's.

There is also a long-term drift to be aware of. Average raw scores on most IQ tests have crept upward across the twentieth century — a phenomenon known as the Flynn effect — which means a score from a test normed in 1980 is not directly comparable to a score from a test normed in 2020. Publishers re-norm their tests periodically to keep the mean anchored at 100, and that re-norming can shift individual scores even when nothing else changes. We cover that story in our piece on the Flynn effect and rising IQ scores.

We offer six IQ test variants on Calculatoris, each built around a specific niche in the comparison table above. Here is a short tour so you can pick the one that matches your reason for being here.

• Quick IQ test — a short, mixed-item assessment designed for curiosity and casual self-checking. The right tool when you want a rough number fast.
• Classical IQ test — a longer, Wechsler-inspired experience that mixes verbal, numerical, and visual items across a broader testing session.
• Mensa-style test — tuned around the abstract reasoning and pattern items that dominate supervised Mensa assessments. Useful practice before a real sitting.
• Raven's-style matrix test — a nonverbal, progressive-matrix assessment in the Raven's tradition, untimed so you can think through each puzzle.
• Raven's-style timed test — the same matrix format with a clock, for readers practicing under time pressure.
• Culture-fair IQ test — a fully visual, language-neutral assessment inspired by the Cattell CFIT tradition for non-native English speakers and cross-cultural use.

The most useful thing you can do before taking an IQ test is to decide what you want from the result. If it is curiosity, a five-minute online test is perfect and anything longer is overkill. If it is practice for a supervised Mensa session, you want an abstract-reasoning-heavy assessment that mirrors the real items. If it is a language-neutral check of fluid reasoning, a Raven's-style or Cattell-inspired visual test is the right fit. And if the result needs to carry real weight — for a diagnosis, a school placement, or a legal decision — no online test substitutes for a licensed psychologist administering a Wechsler or Stanford-Binet in person.

There is no "best" IQ test. There is the test that matches your goal, your budget, and your time. Pick the one that fits your reason for asking, take it in conditions you can trust, and read the score as what it really is — a structured snapshot of how you performed on a specific set of questions on a specific day, not a permanent label on your mind.