Free Visual Memory Test

How many squares can you hold at once?

Visual working memory is the short-lived store that keeps what you have just seen available for the next few seconds. It is small, it is fast, and it is measurable. This test measures it in the most direct way available: a grid of blank squares briefly lights a few of them, the lights vanish, and you have to put them back. Nothing is timed, nothing is verbal, and the only thing that decides how far you get is how much of the pattern survived the moment it disappeared from the screen.

The task starts easy — a three-by-three grid with three lit squares, which almost everyone clears without effort. Every level adds one more lit square, and every third level widens the grid by one row and one column, up to a seven-by-seven board. Two things therefore get harder at once: there is more to remember, and there are more places for it to be. You have three lives. A level is failed when you click a square that was not lit; your score is the highest level you reached before the third mistake.

What the test is actually measuring

Psychologists usually describe visual working memory capacity as being around three to four items for genuinely independent objects. That number sounds far lower than the levels people reach here, and the difference is the whole interesting part of the test. Nobody who reaches level eleven is storing eleven separate square positions. They are storing perhaps three or four groups — a diagonal, a cluster in the corner, a pair on the far edge — and unpacking each group when they click. The capacity limit is real, but it applies to chunks, not to squares, and the skill being measured is largely the skill of forming useful chunks quickly.

This is why the flash is short. Given several seconds, most people would start naming positions — "second row, third column" — and rehearse them silently, which converts a visual task into a verbal one and measures something else entirely. The brief presentation keeps the information in the visual store where it belongs.

How the levels are built

Level one begins with a three-by-three grid and three lit cells. The lit count rises by one every level, so level six has eight lit cells and level twelve has fourteen. The grid steps up on levels four, seven, ten and thirteen, reaching seven-by-seven and staying there. The positions are drawn fresh each time by shuffling every cell index and taking the first few, which means no pattern repeats, no arrangement is impossible, and no level can ever fail to produce enough distinct cells.

Past roughly level fourteen the lit squares cover a third of the board and they begin to form large connected blobs. The task quietly changes character: instead of memorising individual cells it becomes far easier to memorise the blob outlines and the handful of gaps inside them. Players who notice this on their own usually add two or three levels immediately, which is a nice demonstration that a memory score is never purely a memory score — strategy is always part of it. The run ends at level twenty, which almost nobody reaches.

Reading your score

Treat any single run as a rough estimate. Scores bounce around by two or three levels between attempts for the same person on the same day, because one unlucky pattern near your limit ends the run. The average of three or four runs is a far more stable number than your best one.

Getting a better honest score

Look at the whole grid rather than scanning it. Scanning costs time you do not have and tends to leave you with the last corner you looked at and nothing else. Instead, let your eyes rest near the middle and take the pattern in as a single shape. Then, while the grid is blank, click the squares you are most certain about first — a wrong click ends the level, so the order in which you spend your confidence matters.

Practical things also help more than they should. Sit at a comfortable distance, use a real screen rather than a phone held at arm's length, and stop after five or six runs: fatigue on this task shows up quickly and is easy to mistake for a genuine ceiling.

How it differs from our other memory tests

The sequence memory test lights squares one at a time and requires the exact order, so it loads temporal sequencing on top of spatial memory. This test removes order entirely — every lit square appears at the same instant and you may click them in any order you like — which isolates the spatial component much more cleanly. If you want an ability estimate rather than a single capacity measure, the culture fair IQ test covers reasoning rather than memory and is the better place to look.