Free 30-Question Matrix Test

What an untimed matrix test measures that a timed one cannot

Almost every intelligence test you will meet online runs against a clock, and the clock is never a neutral detail. It quietly turns one measurement into two: how well you reason, and how quickly you are willing to commit to an answer. For a large minority of people those two abilities are badly out of step. They see the rule eventually — often completely — but they see it on their own schedule, and a five-minute limit records that as a deficit it is not. This page removes the clock entirely. Thirty matrix puzzles, no timer, no expiry, no penalty of any kind for sitting with a grid until it makes sense. What is left is a much cleaner reading of a single thing: whether you can find a rule you have never been taught.

Why the puzzles are shapes rather than words

Every item is a three-by-three grid of abstract figures with the bottom-right cell missing. No question contains a word, a number to calculate, or any piece of general knowledge, which means your vocabulary, your schooling and your first language have nothing to contribute and nothing to cost you. This is the property psychometricians call culture fairness, and it is the reason matrix formats have survived for eighty years while fashions in verbal testing have come and gone. A test-taker in Lisbon, Seoul or Toronto is looking at exactly the same problem with exactly the same amount of help.

The grids are generated from explicit rules rather than drawn by hand, and the rules fall into four families. In the first, a quantity grows steadily across the columns while shape and shading hold still. In the second, shading cycles through three states across the row while the shape changes from row to row, so two independent rules run at once. In the third, a single figure rotates by a fixed angle per column, which is the family most people find hardest because rotation is easy to see and hard to extrapolate precisely. In the fourth, count is fixed within a row while the shape identity cycles across the columns, forcing you to track a rule that runs horizontally and another that runs vertically.

How the difficulty ramp is built

The thirty items are ordered rather than shuffled. Early puzzles use one rule family with its simplest variant, so the pattern is usually obvious within a few seconds. As you move through the test the variants get less familiar and the families interleave, until the final third combines a rule that runs across the columns with a second rule that runs down the rows. The ramp exists because a flat test wastes your attention: if the tenth item is no harder than the first, everything you learn about your own ceiling comes from a handful of questions at the end. A ramp also gives you a useful private signal. The item where you first stop feeling certain is a rough marker of your own limit, and on an untimed test that marker means something, because you cannot blame it on running out of minutes.

Reading your score

Your raw total is mapped onto the conventional scale where 100 is the population average and fifteen points is one standard deviation. Thirty items is enough to make each question worth roughly three points, so the result is noticeably steadier than a short test — but it is still an estimate produced by a web page, so read it as the centre of a band about eight points wide.

Getting an honest number out of a test with no clock

• Do the whole thing in one sitting if you can. Thirty untimed grids are tiring, and a score collected across three days is measuring three different versions of you.
• Answer every item. Nothing is deducted for a wrong answer, so a blank always costs more than a guess made after eliminating what you can rule out.
• Give a hard grid a real attempt, then move on and return with Back. Fresh eyes on the second pass solve more items than sheer persistence on the first.
• Say the rule out loud before you pick. "Count goes one, two, three; shading is constant" is much harder to fool yourself with than a vague sense of fit.
• Take it once. Repeat attempts drift upward because you have learned the rule families, not because you have become better at finding rules.

How this compares with the other tests here

The quick IQ test puts twenty mixed items behind a five-minute clock and answers a different question: how you perform under pressure. The culture fair IQ test is the closest relative to this one, twenty-five visual items with a generous twenty-minute limit. This page is the version with the pressure removed altogether, which makes it the best of the three if you have ever finished a timed test convinced that the number underrated you. Comparing your untimed score with your timed one is more informative than either number alone: a large gap in favour of this page means your bottleneck is commitment speed rather than reasoning, and that is a fact about you worth knowing.

What this test cannot tell you

A number from a browser is not an assessment. A supervised battery administered by a psychologist controls the room, watches how you approach each problem rather than only what you finally choose, and compares you against a carefully stratified norm sample instead of against whoever happened to visit a website this month. Nothing here should inform a diagnosis, a school placement, or a hiring decision. Use it for curiosity, for practice, and for the honest comparison between timed and untimed reasoning — that is the whole of what it is good for, and it is genuinely useful within those limits.