CompetitiveEasy≈ 8 min

Memory Benchmark

Four short memory tests in under ten minutes — and a profile of how you remember.

By Dmitry Baluev · Updated May 17, 2026

About this test

Four classic memory tasks measure digit span, visual sequence, word recall, and working memory under updating. We turn the four scores into one Memory profile — Polymath, Archivist, Cartographer, Curator, or Dreamer. For best results, take this in a quiet space with minimal distractions. On mobile, use the on-screen number pad we provide; don't switch apps mid-test, since browsers may interrupt the timer.

Start the battery
#memory#working-memory#digit-span#n-back#battery#4-tests

What your result could be

Every possible result for this test, and what each one means.

Polymath — Cross-domain memory — no weak axis.
Your composite sits in the top band. Polymath is the name for that level: the strength is not one drilled mode but a broad capacity working across several formats at once. Cross-domain memory makes you the person who can hold a meeting agenda in mind while also remembering the room layout, the names of the people present, and the gist of a side conversation that happened during the break. The archetype name comes from the overall level rather than any single scale — the shape of your own profile is worked out from your four results and set out below. Where this shines: assimilating new material across formats — text, diagrams, sequences, melody, spatial arrangements — without disproportionate effort in any one mode. This profile shows up in academic high performers, polyglot language learners, professional musicians who also read scores, scientists who maintain both written and mental indexes of their field, generalist clinicians who hold detailed patient histories alongside spatial findings, and senior engineers whose mental model spans architecture, code, and conversation history. The cross-format capacity scales the kinds of roles you can hold simultaneously without dropping context. The blind-spot worth knowing: people whose memory works as easily as yours systematically underestimate the effort other people put into memorizing material in formats that come for free to them. The trap is assuming everyone's memory works like yours does — and reading collaborators' slower assimilation of meeting notes, spec documents, or conversation history as inattention or low engagement. It's usually neither. It's a different memory shape doing the best it can with different equipment. The growth direction is awareness rather than further capacity. The Polymath profile already does more lifting than it gets credit for; the practice worth building is making your memory legible to people who don't share it — written summaries, follow-up notes, explicit recaps — so the team can run on shared external memory instead of your private internal index.
Archivist — Strong organizing recall — builds long mental indexes.
Your composite sits in the upper band. Archivist is the name for the strategy that usually carries a level like this one: fidelity over breadth. The pattern the name points at reads as mental archive: long indexed structures get built in your head — chronologies, taxonomies, ordered lists, phone-number-style sequences — and they come back with high accuracy when queried. The archetype name comes from the overall level rather than any single scale — the shape of your own profile is worked out from your four results and set out below. Where this shines: anywhere fidelity matters more than synthesis. People who work with you notice you can recite specific quotes from meetings, recall the order of items on past agendas, and reconstruct conversations with linguistic precision. This profile shows up in writers, editors, oral historians, legal researchers, taxonomists, librarians, classical-music performers, and programmers who hold large symbol tables in mind. That kind of fidelity compounds over time: archives feed each other, and a long-running mental index tends to grow more useful, not less, as the years stack up. The blind-spot worth knowing: the same fidelity that makes you a careful archivist can shade into treating every conversation as something to be catalogued. Sometimes people want to be heard, not transcribed. Watch for the version where a friend's offhand remark gets pulled out as evidence in a disagreement six months later — accurate recall is not always the kindest move. The growth direction is the gist-vs-fidelity tradeoff in reverse. The Archivist's habit is to encode in detail, even when detail isn't what the situation needs. Practicing the discipline of summary — what is this fundamentally about, in one sentence — uses a complementary capacity that doesn't naturally come with the same equipment.
Cartographer — Builds memory like maps — integrates spatial patterns.
Your composite sits in the middle band. Cartographer is the name for the approach that usually rescues a level like this one: leaning on structure and relations instead of trying to hold everything as a list. The pattern the name points at reads as mental cartography: the world gets encoded through spatial relations rather than through abstract lists, and recall flows back through where things were, not which order they came in. The archetype name comes from the overall level rather than any single scale — the shape of your own profile is worked out from your four results and set out below. Where this shines: navigation in unfamiliar physical and informational spaces. You find the section of a long document by remembering it was on the left side near the bottom of a page about three-quarters in. You trace organizational structures across a company you joined recently. You walk through a building once and rebuild the floor plan from memory. Real-world: this profile shows up in architects, urban planners, surgeons, geographers, choreographers, infrastructure engineers, video-game level designers, and field biologists. Software engineers of this cast tend to remember directory trees and file relationships rather than function names verbatim. The blind-spot worth knowing: if Chain — the digit-span sub-test — came out low for you, take it as a real pattern rather than a fluke of one round. Phone numbers, account numbers, verification codes, and ordered numerical strings in general then take more deliberate effort than spatial recall would predict. Voice memos, password managers, and write-down-immediately habits cover the gap without requiring you to retrain the underlying capacity. The growth direction is to lean into the spatial encoding instead of treating it as a partial profile. Method-of-loci techniques, spatial diagrams instead of bullet lists, sketched maps of complex projects — these aren't workarounds; they're using the mode your memory is already organized around.
Curator — Selective recall — keeps what matters, discards the rest.
Your composite sits in the lower-middle band. Curator is the name for reading that level as selective curation rather than as low-capacity storage: mental bandwidth is allocated narrowly but well. You don't try to remember everything; you keep what your attention judged important in the moment and let the rest fall away cleanly. This is a different memory strategy from the broad-storage profiles (Polymath, Archivist) — closer to a museum curator with a small, well-chosen permanent collection than a librarian with an exhaustive catalogue. The archetype name comes from the overall level rather than any single scale — the shape of your own profile is worked out from your four results and set out below. Where this shines: holding a tight working set under pressure while ignoring noise that doesn't matter. Editors who cut ruthlessly, product managers who keep three priorities in view while a hundred distractions arrive each week, chess players who track key pieces while letting irrelevant positions blur, writers who release drafts rather than archive every variant — these are Curator-shaped roles. If your own peak landed on Twin, that is the same updating capacity that supports live conversation, dynamic problem-solving, and tasks where the right answer keeps shifting. The blind-spot worth knowing: people who archive everything will sometimes read your selectivity as forgetfulness or as 'not paying attention'. What looks like 'I don't remember' is often 'I deliberately released it' — but the social signal can land the same way. In collaborative work, a brief written summary after meetings closes the gap without forcing your mind to become an Archivist it isn't. The growth direction is to keep trusting the selection process rather than over-correcting toward broad storage. The Polymath profile is not the goal state, and the strategies that would push your scores higher (rehearsal drills, mnemonic systems) often work by displacing the very curation skill that's already serving you.
Dreamer — Lives in present and imagination — doesn't rely on accumulated memory.
Your composite sits below the published anchor curves, and the honest reading is that controlled list-recall in a noisy browser environment is not where your mind allocates its bandwidth. 'Dreamer' is the word for someone whose mental life is oriented toward the present moment and the imagined future rather than the indexed past. The nervous system spends less effort accumulating and retrieving stored material, and more on in-the-moment perception, association, and creative recombination. This is a different cognitive strategy from the broad-archive profiles above — not a deficit, not something to fix. Where this shines: synthetic insight that comes from letting material settle quietly without forced retention; comfort with ambiguity and partially-formed thoughts; perceptual freshness in domains where over-indexing the past gets in the way. People with this profile show up disproportionately among artists, fiction writers, improvisational performers, designers, and contemplative practitioners. They also show up in scientists and engineers whose contribution is reframing a stuck problem rather than recalling the citation chain that led to it. The blind-spot worth knowing: low list-recall on a timed browser test correlates loosely with the experience of feeling 'forgetful' in conversations, meetings, and study contexts that reward verbatim playback. External memory aids — written notes, calendars, voice memos, photos of whiteboards, structured documents — are legitimate extensions of cognition, not workarounds for something missing. Treat them as the equivalent of a writer's notebook, not as a crutch. The growth direction is sideways, not upward. The memory-test score is one narrow measurement of one narrow capability. It does not measure intelligence, creativity, social attunement, emotional depth, intuitive judgment, or synthetic capacity. The trap is letting one number define a self-story — your mind has many more dimensions than this battery sampled.

Frequently asked

How accurate is this test?

Treat it as a playful snapshot, not a clinical measurement. Memory Benchmark is four short tasks in a browser — useful for self-discovery and comparison, not for diagnosis. Memory varies hour-to-hour with sleep, stress, mood, caffeine, time of day, and many other factors. For a more stable picture, run the battery two or three times across different days and look at the pattern, not the single number.

What do the four sub-tests actually measure?

Chain (digit span) tests how many digits you can hold in working memory in order — the classic Miller 7±2 capacity. Trail (visual sequence) is the visuospatial counterpart, based on Corsi 1972. Wordlist tests verbal recognition memory with a short interference delay (Sternberg 1966 / Brown-Peterson 1958). Twin (N-back, n=2) tests working memory under continuous updating (Kirchner 1958). Together they cover all three components of Baddeley & Hitch's working-memory model.

Why is the mobile experience different from desktop?

We custom-built an on-screen number pad for the Chain sub-test specifically because the dominant memory-test site triggers the native OS keyboard, which covers half the screen during recall. That's the single biggest UX flaw in the category, and we close it. Other sub-tests are tap-based and work the same on both platforms. Avoid switching tabs mid-test — browsers may pause timers and void the round.

What if I scored roughly the same on all four sub-tests?

That's a balanced profile — either the Polymath shape (balanced-high) or the Dreamer shape (balanced-low) depending on where the composite sits. A flat profile is informative on its own: it tells you your memory works as a general-purpose system rather than as a set of specialized subsystems, and it predicts how you'll handle novel learning situations across formats.

Can my memory improve with practice?

Direct practice on a specific memory task yields some improvement that stabilizes after a few sessions and is largely task-specific (you get better at digit span without getting better at face recognition). Sleep, hydration, and ambient stress have much larger same-session effects. We do not claim that this test 'trains your brain' or improves real-world memory, and you should be skeptical of products that do — the science on transfer effects from memory-training apps to everyday cognition is mixed at best.

Why don't you call this a Memory IQ or Brain Age?

Because those framings overstate what a five-minute browser test can measure. 'Brain age' specifically is adjacent to language that the US Federal Trade Commission has policed in this product category. We report your raw scores and a percentile against published academic norms — that's honest. The archetype is a shareable interpretive frame; it isn't a diagnosis.

Why use a recognition test for Wordlist instead of free recall?

Free recall (saying or typing the words you remember) is fairly described as the gold-standard test in clinical neuropsychology — but it's a mobile UX nightmare (free typing on phones is slow and frustrating). Recognition memory (tapping seen words from a pool of 30) gives a robust signal-detection score (hits minus false alarms) that doesn't penalize slow typing. The Sternberg 1966 paradigm we use is well-validated and academically standard.

How does this compare to clinical memory tests like WAIS Digit Span or Cambridge Brain Sciences?

The four sub-tests in Memory Benchmark are built from the same public-domain paradigms that underlie those clinical instruments (Wechsler-lineage digit span, Corsi block-tapping, Sternberg recognition, Kirchner n-back). However, this is not a clinical test — clinical instruments come with standardized administration manuals, normed against specific populations, and interpreted by trained psychologists. Memory Benchmark is a playful self-discovery battery; we use the same paradigm shapes but our own UI, our own item generation, and our own scoring presentation. Treat your result as a conversation-starter, not as clinical assessment. If you have real concerns about memory function, see a qualified clinician.

Related tests

≈ 4 minEasyCompetitive

Mind Age

Four quick speed-and-memory games turn into one number: your cognitive age — just for fun.

Take the test
≈ 2 minMediumCompetitive

The Chimp Test

Can you beat a chimpanzee at working memory?

Take the test
≈ 5 minEasyCompetitive

Reflex Profile

Five reaction tests in five minutes — and a profile of how you respond.

Take the test
All Skills tests →