People searching for the man vs chimp website are almost always looking for one specific task: numerals appear scattered on a screen, they hide, and you touch them in ascending order from memory. It is playable here with no account, and the version below is an independent recreation.

Benchmento is not affiliated with the Kyoto University Primate Research Institute, with any of its researchers, or with any other site hosting a version of this task.

Numbers held
Numbers heldStanding
4 – 5Getting started
6 – 8Below to approaching average
9 – 10Around average
11 – 13Good
14 – 16Very good
17+Exceptional

What the Kyoto work actually showed

The research that produced the viral clips was done at Kyoto University with chimpanzees who had been taught the ordering of Arabic numerals over a long period. The best known participant, Ayumu, was tested on a task where numerals flashed briefly and then turned into blank squares, and he was able to touch them in the correct order at exposures short enough that most human participants performed poorly.

The claim that circulated afterwards was much wider than the finding. The result was about a specific, heavily trained task at very short exposure times, not a general statement that chimpanzees have better memories than people. Later work argued that humans given comparable practice narrow the gap substantially, and the interpretation is still debated. A browser version cannot settle any of that.

How this version differs

The version here is not a reproduction of the laboratory protocol and should not be read as one. In the original, exposure was controlled to fractions of a second. Here you can look at the numerals for as long as you like, and they hide when you click the first one. That makes it a test of how much you can encode when you choose to commit, rather than of how much you take in from a fixed glance.

You are also using a mouse or a touchscreen rather than a trained touch panel, and you have not spent years learning the task. All of that makes the numbers here your own benchmark rather than something to hold against a chimpanzee.

What counts as a good score

Around nine or ten is average. Eleven to thirteen is good, fourteen to sixteen is very good, and seventeen or more is exceptional. Above roughly fifteen most people are no longer holding individual positions in memory; they are encoding a path or a shape across the screen, which is a different and very effective strategy.

Why people search for it

The task went viral through short clips of a chimpanzee clearing a screen of numerals faster than the human sitting beside it, and those clips are still circulating years later. Most people arrive looking for the version they saw rather than the paper behind it, which is why the phrase they type is usually the man vs chimp website rather than anything a researcher would call it.

If you want the closely related span tasks, the sequence, number and visual memory tests measure the same underlying capacity in different formats and take about the same time to play.

It is not an intelligence test

This measures one narrow slice of visual working memory. It produces no IQ score, it is not a psychometric instrument, and performance improves with practice and strategy in a way that intelligence measures are specifically designed not to. Treat the number as a benchmark against yourself.

Take the chimp testOr the sequence memory test · number memory test · visual memory test.

Common questions

Is this the man vs chimp website?

This is an independent browser version of the numeral task that people search for under that name. It is not affiliated with the Kyoto University Primate Research Institute or with any other site hosting the task.

Did chimpanzees really beat humans at memory?

On one specific task, at very short exposure times, trained chimpanzees outperformed the human participants tested. The wider claim that chimpanzees have better memories in general is not what the research showed, and later work argued that practised humans close much of the gap.

What is a good score?

Around nine or ten is average, eleven to thirteen is good, and seventeen or more is exceptional. Very high scores usually come from encoding a path across the screen rather than holding each position separately.

Is the chimp test an IQ test?

No. It measures one narrow component of visual working memory and produces no IQ score. Performance improves with practice and strategy, which is not how an intelligence measure is supposed to behave.