Direct answer
The average Chimp Test score is roughly level 6. Most people fail somewhere in the 4 to 6 range, where the grid grows past the span of visual working memory you can hold under time pressure. Reaching level 7 is solidly above average, level 8 is excellent and lands you in roughly the top 5% of players, and level 9 or higher is rare for an untrained human.
The Chimp Test starts at 4 numbered tiles and adds one tile each round. Your score is the highest level you clear, so a "score of 6" means you correctly recalled a 6-tile board after the numbers were masked. That is why scores cluster as small whole numbers rather than spreading across a wide range like a reaction-time test in milliseconds.
Why some sources say 8-12 (and why that's misleading)
Search around and you'll find two contradictory claims: that the average is about 6, and that it's 8 to 12. Both can be true because they measure different populations under different conditions, and the higher figure rarely reflects a typical first attempt.
The 8-12 range usually comes from three sources of inflation. First, self-reported "good" scores from forums and comment threads are survivor-biased: people who hit level 4 don't post about it, people who hit level 11 do. Second, repeated attempts matter enormously. The first time you take the test you're learning the interface; by the tenth attempt you've developed a chunking strategy and your ceiling rises several levels. Third, implementations differ. Number visibility time, grid size, and mouse versus touch all shift the difficulty, so a score on one clone isn't comparable to a score on another.
- Average for a genuine first attempt: about level 5-6
- Above average / good: level 7
- Excellent (top ~5%): level 8
- Rare for untrained humans: level 9+
- "8-12" figures usually reflect practiced, self-selected, or different-implementation scores
Chimp Test score benchmarks
Use this as a rough ladder rather than a precise percentile chart. Exact distributions aren't published by the major test sites, so these bands reflect the commonly cited consensus that most players stall at 4-6 and that level 8 represents top-tier human performance.
| Level reached | Interpretation | Roughly where you stand |
|---|---|---|
| 4-5 | Below to around average | Where most first attempts end |
| 6 | Average | The typical ceiling |
| 7 | Above average | Better than most players |
| 8 | Excellent | Approx. top 5% |
| 9-10 | Exceptional | Rare for untrained humans |
| 11+ | Outlier / heavily practiced | Usually many attempts or a different implementation |
The chimpanzee research the test is named after
The Chimp Test references a real and frequently misreported experiment. In 2007, Sana Inoue and Tetsuro Matsuzawa at Kyoto University's Primate Research Institute published "Working memory of numerals in chimpanzees" in Current Biology. A young chimpanzee named Ayumu was trained to touch the numerals 1 through 9 on a screen in ascending order. In the hardest version, the numbers flashed for as little as 210 milliseconds before being masked by white squares, and Ayumu had to tap the hidden positions from memory.
Ayumu held roughly 80% accuracy on five-numeral trials even at 210 ms, and crucially his accuracy barely dropped as the display time shortened. The human subjects in the same apparatus got worse as the flash got faster. That single comparison produced the famous headline that young chimps beat university students at memory, and it fed the cognitive trade-off hypothesis: the idea that humans may have given up some rapid visuospatial memory as we evolved language.
The Chimp Test you play online is a simplified, self-paced version of this paradigm. It shares the core mechanic (memorize numbered positions, recall them after they're hidden) but not the millisecond timing control, so your level is not directly comparable to Ayumu's accuracy figures.
The part the headlines usually leave out
The "chimps beat humans" story is real but incomplete. Ayumu had years of practice on the task; the human comparison subjects had almost none. When Cook and Wilson (2010) trained two undergraduates properly, giving them around 6,000 practice trials, the humans hit 94% and 96% accuracy at the same 210 ms, five-numeral condition, significantly outperforming the chimpanzees.
Their conclusion was blunt: there's no evidence for a superior or qualitatively different spatial-memory system in chimpanzees once you equalize practice. The chimp advantage in the original study largely reflected a training gap, not a species gap. So if your Chimp Test score climbs after a dozen attempts, that's the same mechanism at work, and it's exactly what you'd expect. The test rewards practiced chunking strategy as much as raw memory span.
How to actually improve your score
Because the bottleneck is visual working memory under time pressure, the gains come from encoding the board faster, not from staring longer. Chunk the layout into spatial groups, read positions in a consistent scan order, and rehearse the sequence as a path rather than nine separate dots.
- Group tiles into 2-3 spatial clusters instead of memorizing each one
- Lock in the first few positions instantly, then fill in the rest
- Take it rested; visual working memory is sensitive to fatigue
- Expect the first few runs to undersell you, then plateau at your true ceiling
Frequently asked questions
What is the average Chimp Test score? About level 6. Most players stall somewhere between level 4 and 6 on a genuine first attempt, because that's where the number of tiles exceeds what you can hold in visual working memory after they're masked.
Is a Chimp Test score of 8 good? Yes. Level 8 is excellent and lands you in roughly the top 5% of players. Level 7 is already above average, and level 9 or higher is rare for someone who hasn't practiced the task repeatedly.
Why do some sites say the average is 8 to 12? Those figures usually reflect practiced or self-reported scores, not first attempts. People who post their results skew high, repeated attempts raise your ceiling several levels, and different test implementations vary in difficulty, so the numbers aren't comparable.
Did chimpanzees really beat humans on this test? In the 2007 Inoue and Matsuzawa study the chimp Ayumu outperformed humans, but the humans had little practice and Ayumu had years. A 2010 follow-up by Cook and Wilson trained humans properly and they scored 94-96%, beating the chimps. The original gap was mostly a training gap.
Does the Chimp Test measure IQ? No. It measures visual-spatial working memory under time pressure, one narrow component of cognition. It's not a validated IQ test, and a high or low score doesn't translate directly to general intelligence.
Related on Deadline
take the Chimp Test · Visual Memory Test · all memory tests · Sequence Memory Test
Sources and notes
- https://www.cell.com/current-biology/fulltext/S0960-9822(07)02088-X
- https://www.cell.com/current-biology/pdf/S0960-9822(07)02088-X.pdf
- https://link.springer.com/article/10.3758/PBR.17.4.599
- https://escholarship.org/content/qt5d06v437/qt5d06v437_noSplash_93a9354c784d64eed437b77814124c60.pdf
- https://en.wikipedia.org/wiki/Cognitive_tradeoff_hypothesis
- https://blog.jcx.au/posts/humans-1-chimps-0-correcting-the-record
- https://www.nationalgeographic.com/science/article/chimps-trump-university-students-at-memory-task
- https://www.measurehuman.com/guides/average-chimp-test