AbilityBenchAdmission-test practice
Practice IQ test in the Mensa admission format
Twenty-four generated figure items — matrices, series and classifications — on one twelve-minute clock, scored as a raw number correct the moment the clock stops. It is practice for the shape of a supervised admission test, not the test: no unsupervised online test qualifies anyone, and no IQ figure is printed, because a browser has no standardization sample to rank you against. What you get is the score, the odds of reaching it by guessing, and the rule behind every item you missed.
- 24 items in 12 minutes
- Six options per item
- Raw score, not an IQ
24 figure items in 12 minutes — matrices, series and classifications, six options each, generated in this tab from a seed you can copy afterwards. The set is ordered easiest first, so the clock takes the hardest items away rather than a random handful.
The run, in full
- Items
- 24 scored — eight matrices, eight series, eight classifications — plus 3 practice items excluded from every figure
- Clock
- 12 minutes across the whole scored set, about 30 seconds an item, and it stops while this tab is in the background
- Options
- 6 per item, one correct, so a guess is right once in 6
- Answering
- keys 1 to 6, the number pad, or a tap on the option itself
- One pass
- each item is answered once, there is no review pass, and a wrong answer costs exactly what a blank does
- Reported
- items correct out of 24, split by item type, with an interval and the odds against guessing — never an IQ
The practice items are one of each kind, untimed, and each one names the rule it was built from afterwards — including the rule you just used without putting a name to it. Nothing in them reaches the score, and the clock does not start until the timed set does.
One percentile, three different numbers
The admission criterion is a rank, not a score: the 98th percentile of the general population, per Mensa International's published admission criterion. A rank only becomes a number once it is put on a scale, and the IQ scales in circulation share a mean of 100 but not a spread — so the same rank comes out as three numbers. The right-hand column is computed on this page from the percentile and the standard deviation, on the normal curve the scales are built on.
| Scale | Standard deviation | Score at that percentile |
|---|---|---|
| Wechsler (WAIS, WISC)The Wechsler scales are constructed with a mean of 100 and a standard deviation of 15 | 15 | 130.8 |
| Stanford-Binet Form L-MForm L-M was constructed with a standard deviation of 16 | 16 | 132.9 |
| CattellThe Cattell scales are constructed with a standard deviation of 24 | 24 | 149.3 |
Nobody is disagreeing with anybody when two of those figures are quoted for one criterion; they are one rank in two units. It is also why an IQ figure with no scale named beside it cannot be checked, and why the second question to ask of any such number is when the test was normed: performance drifts upward against fixed norms by ≈ 3 IQ points a decade (Flynn (1987), Massive IQ gains in 14 nations: What IQ tests really measure, Psychological Bulletin), fastest on the abstract figure items free online tests copy, so an undated table flatters everyone it scores.
Before a supervised sitting
Can a score from this page qualify me for Mensa?
No, and neither can any other online test. Qualification needs a supervised sitting — a proctor, an approved instrument, timing that is the same for everyone, and a result reported by the tester rather than by the person tested — or evidence of a score from such a sitting elsewhere, on a test the organization accepts. An unwatched score is a claim about yourself. What an online set can do is take the surprise out of the formats before the day, which is what this one is for.
Should I guess on an item I cannot see through?
Yes. The score is the number right with nothing subtracted, so a wrong answer and a blank cost the same, and with six options a blind guess lands one time in six — about four items if you guessed a whole set. An item you have narrowed to two is worth a good deal more than that. The result prints the exact probability of your whole score arising from guessing, which is the honest way to read a set this short.
What happens if I switch tabs, or the clock runs out mid-item?
The item on screen is dropped and not asked again, and the clock stops while the tab is in the background, so the time away is not charged to you. Asking the item a second time would measure whether you remembered it. If the twelve minutes run out, the items you never reached count as unanswered, and they are the hardest ones, because the set runs easiest first — the clock takes the top of the set rather than a random handful.
Generated items, and the format left out
The published item banks are copyrighted and kept out of circulation, because an item somebody has seen tests recall rather than reasoning. So every item here is generated from construction rules — a value held constant along a row, a value fixed down a column, a value distributed so each of three appears once in every row and every column — of the kind Carpenter, Just & Shell (1990), What one intelligence test measures: A theoretical account of the processing in the Raven Progressive Matrices Test, Psychological Review found behind a published matrix test. No two attributes of a matrix share a rule, which keeps the nine cells nine different drawings; an item where two cells coincide usually has a second answer someone can defend. No color is used either: the fills are outline, gray and solid, so the set reads the same to someone with a red-green deficiency.
One format the culture-fair batteries use is deliberately missing: the topology items, where a dot has to fit inside one shape and outside another. A generated version of those is only unambiguous when the geometry is checked by hand, and an item with two defensible answers is worse than no item. If the grids were the part you wanted more of, the Raven’s-format matrices take the clock away and report which construction rule you lost; the number half of an admission battery is closer to the math aptitude test.
Not affiliated with, endorsed by, or connected to the owner of Mensa. Mensa is a trademark of its owner and is used here only to name the assessment this page prepares for. The questions on this page are our own: no part of the published test is reproduced, and a score here is not comparable to a score from the real instrument.
This is a measurement exercise, not a clinical assessment. It reports what you did on this page against a stated reference and nothing more — it cannot establish giftedness, a learning difficulty, or eligibility for anything. Only a qualified professional, working with more than a browser, can make that judgment.