10 instruments · free · nothing leaves your browser
Human benchmark tests for reaction time, typing speed, color vision and hearing
Start with the reaction test below — click the panel, wait for it to change, and hit it as fast as you can. Five trials take about twenty seconds and report your median in milliseconds against a reference distribution. Every other instrument on the site works the same way: a real measurement, the formula shown, no sign-up and no data leaving your device.
Every instrument on the bench
Each page is a working test plus the method behind it. Nothing is gated, nothing asks for an email, and no advert ever sits between you and a result.
- net WPM · accuracy %typing testTimed typing test with live caret, per-key error heatmap and net WPM computed from your own character counts.
- gross WPM · net WPMwords per minute typing testThe WPM formula written out with your numbers substituted in, plus what each score means by profession.
- WPM · CPM · accuracycheck your typing speedA 25-word sprint that gives a usable speed reading in under a minute, plus the four things that distort it.
- your percentileaverage typing speed by percentileThe full distribution table — where 40 WPM actually sits, and what the 25th, 50th, 75th and 95th percentiles are.
- type · rough severitycolor blind testPseudoisochromatic plates generated live from cone-confusion lines. Screening indicator, not a diagnosis.
- protan / deutan / tritancolor vision deficiency testThe same plates, graded by confusion-line strength, so protan, deutan and tritan are separated rather than lumped.
- median ms · percentilereaction time testFive trials on performance.now() with false-start detection, then median, mean and best against published norms.
- ms/target · accuracy %aim trainerThirty targets, measured time-to-target and click accuracy — the pointing half of reaction speed.
- highest audible Hzhearing range testWeb Audio pure-tone sweep from 20 Hz to 20 kHz, per ear, with calibration first. Screening indicator, not a diagnosis.
- cm · fractional inchesonline rulerCalibrated against a credit card (85.60 mm by ISO/IEC 7810), so centimeters are real centimeters on any screen.
What each number actually measures
A benchmark is only useful if you know what is inside it. These are the four honest caveats that matter most, one per instrument family.
- Reaction time measured in a browser includes the display and the input chain, not just you. A 60 Hz screen holds each frame for 16.7 ms, so the stimulus can wait up to a full frame before your eye sees it, and the click has to be sampled and dispatched. Expect 20–40 ms of that reading to belong to the hardware.
- Typing speed depends on the text. A test drawn from the most frequent English words produces a higher number than one with punctuation and proper nouns, so a score is only comparable against a test with the same prompt style. Ours states the word list on the page.
- Color vision plates are generated from cone-confusion geometry and rendered on a display nobody has profiled. Night Shift, an ambient-light sensor or a cheap panel each move the colors enough to matter. That is why it is a screening indicator and never a diagnosis.
- Hearing in a browser has no calibrated sound pressure level at all. What can be measured honestly is the highest frequency you still detect at a fixed comfortable volume, which is genuinely informative because age-related loss starts at the top of the range.
Search demand for these tests
Every route on this site was chosen against measured US search volume rather than a guess. The figures below are Google Keyword Planner, July 2026 — they are why these nine instruments exist and, for example, why there is no “memory test” here yet.
| Instrument | Primary keyword | Monthly searches |
|---|---|---|
| typing test | typing test | 673,000 |
| words per minute typing test | words per minute typing test | 135,000 |
| check your typing speed | check typing speed | 135,000 |
| average typing speed by percentile | average typing speed | 33,100 |
| color blind test | color blind test | 201,000 |
| color vision deficiency test | color vision deficiency test | 201,000 |
| reaction time test | reaction time test | 135,000 |
| aim trainer | aim trainer | 33,100 |
| hearing range test | hearing test | 27,100 |
| online ruler | online ruler | 22,200 |
Questions about these tests
What is a human benchmark test?
A short, repeatable task that produces a number you can compare against a reference distribution — reaction time in milliseconds, typing speed in words per minute, the highest audible frequency in hertz. The value is in the repeatability: one run tells you very little, five runs on different days tells you where you actually sit.
Are the results accurate?
The arithmetic is exact and every formula is shown on its page. The measurement is only as good as the hardware: reaction time includes display and input latency, color plates depend on an uncalibrated screen, and a browser cannot know the sound pressure level at your eardrum. Each page states what its own reading includes and what it cannot.
Do you store my results?
No. Every test runs entirely in your browser. Personal bests are written to localStorage on your own device, which means they survive a reload and never leave the machine. There are no accounts and nothing to sign up for.
Can these tests diagnose anything?
No. The color vision and hearing pages are screening indicators and say so prominently. They cannot diagnose a condition and they cannot rule one out. If a result concerns you, or the stakes are real — a job, a license, a change you have noticed — see an optometrist, an audiologist or a doctor.
Which test should I start with?
The reaction test on this page: it takes about twenty seconds and needs nothing but a mouse or a fingertip. If you are here for a typing figure, go straight to the one-minute typing test, which reports net WPM with the arithmetic expanded.
Do these work on a phone?
Reaction time, aim, color vision and the ruler work well on a phone. The typing test works but a touch keyboard costs most people 20 to 30 WPM against a physical one, so treat a phone score as its own category. The hearing sweep needs wired headphones — phone speakers roll off the top octave entirely.
Testing the device instead of the person?
These instruments measure a human. If what you actually need is a measurement of the hardware — scroll behavior, touch sampling, refresh rate, input latency — that lives on scrolltest.io. And if you want to see what a colorblind viewer sees rather than screen your own vision, the color blindness simulator at hextorgb.io applies the same cone-space transforms used by the color blind test here, but in the opposite direction.
HumanBenchTests — Instruments for measuring human performance, in the browser.