What Is on the Mensa Test? Question Formats and Difficulty Explained

The Mensa admission test centers on non-verbal problems about patterns in figures, and the qualifying threshold is the top 2%. There are no knowledge questions. This article explains the format and difficulty within the limits of what is publicly documented, along with the limits of preparation.

First: the actual items are not public

To be straightforward about it, the items used on Mensa's admission test are not published. Keeping the test valid requires that real items not circulate in advance, so anything offered online as a 'Mensa past paper' is best treated as a practice item in a similar format rather than an official one.

The nature of the test, however, is public. What follows describes format and threshold in general terms; it does not reproduce any individual item.

Format: non-verbal problems about patterns in figures

Mensa uses standardized intelligence tests in a non-verbal format. The core question types are:

• Matrix reasoning: choose the figure that fills the blank cell in a 3x3 grid (Raven's-matrices style)
• Series completion: choose what comes next in a sequence of figures changing by a fixed rule
• Classification: choose the figure that does not follow the shared rule

None of these require language knowledge, arithmetic, or general education. The design is 'culture-fair,' meaning it is relatively insensitive to native language and schooling — the same principle this site's test follows (see our methodology). Note that the specific instrument varies by country and year, so the details of the format are not uniform.

The difficulty benchmark is the top 2%

Mensa's admission criterion is scoring in the top 2% on a standardized intelligence test. On a normal distribution with a mean of 100 and a standard deviation of 15, the top 2% corresponds to IQ 130 and above (see the percentage for IQ 130).

The top 2% is about 1 in 50 people. In terms of how it feels, the test opens with items most people can solve and then increases the number of rules operating at once. As Carpenter, Just, and Shell (1990) showed, difficulty in matrix reasoning is largely explained by how many rules hold simultaneously within a single item. At the top-2% level, picturing three or more concurrent rules is a reasonable mental model.

What preparation improves, and what it does not

The effect of preparation splits cleanly:

• Improves: familiarity with the format (fewer points lost to first-encounter hesitation), time management
• Resists improvement: fluid reasoning itself — the ability to discover an unfamiliar rule

Repeating items in the same format is known to raise scores by a few points as a practice effect, but this is not an increase in ability; it is moving closer to your true baseline. We cover this in does practicing for IQ tests help?. So walking through the formats once is sensible, while cramming for months to clear the threshold is a poor use of effort.

Gauge your position before you sit the test

Because the admission test costs money and runs on fixed dates, knowing roughly where you stand beforehand makes the decision easier. The application process is summarized in how to join Mensa.

BrainRank's free IQ test is built around non-verbal figure reasoning like the Mensa test, running 20 questions in about 10 minutes and using item response theory (IRT) scoring to produce an estimated IQ on the mean-100, SD-15 scale. It is an indication rather than an official determination, but it is a starting point for measuring your distance from the top-2% line.