What does this test actually measure?

Two browser-observed latencies: one signal to one response, then one signal to a correct two-way choice. The report also records variation, errors, false starts, and input method.

Does a lower number mean I am better?

Not by itself. Device latency, input method, anticipation, and accuracy all affect a browser measurement. Compare repeated runs on the same setup and read the raw trials with the errors.

Why does the test require ten trials and not one?

One observation is too fragile. Five trials per task allow a median and spread, but this still remains a lightweight prototype rather than a validated laboratory measure.

Is mobile or desktop more accurate?

Neither is universally more accurate here. Hardware, browser, display, and input path differ. Keep the device and input method constant when comparing your own runs.

Can this report explain why I was slower?

No. It can show that this run was slower or noisier under current conditions, but it cannot identify fatigue, health, age, attention, or any other cause.

Atlas/Mind/C.01 — Reaction time

Reaction time test

Plate C.01 · 2 min · 10 trials.

How fastare you?

Ten timed trials in your browser: tap on the signal, then pick the right key under pressure. You get your reaction time in milliseconds, how much choosing slows you down, and every trial on record.

Start the testno signup · no email · 2 min

§ 0 · At a glance

Reaction time test: free on Mesonar, 10 timed trials after two warm-ups, about 2 minutes. No signup, no email, no app.
5 possible results: fast and steady, quick but costly on choices, measured start, slower on this run, or a retest flag when the run is too noisy.
Half the trials are simple (react to one signal) and half are choice (press the matching key); the report shows every timing, both medians in milliseconds, the spread, the cost of choosing and an error check.
It measures reaction speed on your own device and browser; screen refresh, input lag and fatigue all move the numbers.
Available in 10 languages: English, 简体中文, 繁體中文, 日本語, 한국어, Español, Português, Deutsch, Français and Русский.
Facts last reviewed October 2026.

§ I · Two tasks, one deviceSimple · Choice

Reflex first.
Choice second.

Simple reaction records the delay from one visual signal to one response. Choice reaction adds a correct left-or-right selection. Their median difference is reported as choice cost.

Simple reaction

5 measured trials after two warm-ups. Wait for the signal, then respond. False starts remain part of the run-quality check.

Lower median< 200 ms

Compact spread< 250 ms

This run< 300 ms

Choice reaction

5 measured trials. A red or blue signal appears; choose the matching side. Wrong choices remain visible and the trial is repeated.

Lower median< 350 ms

Compact spread< 420 ms

This run< 500 ms

§ II · Four descriptions plus a retest stateFast & steady · Choice cost · Measured start · Slower run · Retest

quick medians with a narrow trial-to-trial spread

Fast and Steady

The signal arrived quickly, and your timing stayed compact.

the first response was quick; choosing added more time

Quick Signal, Costly Choice

Your first signal moved fast. Choice made the gap visible.

the start took longer; choosing added comparatively little

Measured Start, Small Choice Cost

The first move was measured. The second decision stayed close.

both medians were higher under the current conditions

Slower on This Run

This run was slower. The reason remains unknown.

§ III · Frequently asked5 q.

What does this test actually measure?: Two browser-observed latencies: one signal to one response, then one signal to a correct two-way choice. The report also records variation, errors, false starts, and input method.
Does a lower number mean I am better?: Not by itself. Device latency, input method, anticipation, and accuracy all affect a browser measurement. Compare repeated runs on the same setup and read the raw trials with the errors.
Why does the test require ten trials and not one?: One observation is too fragile. Five trials per task allow a median and spread, but this still remains a lightweight prototype rather than a validated laboratory measure.
Is mobile or desktop more accurate?: Neither is universally more accurate here. Hardware, browser, display, and input path differ. Keep the device and input method constant when comparing your own runs.
Can this report explain why I was slower?: No. It can show that this run was slower or noisier under current conditions, but it cannot identify fatigue, health, age, attention, or any other cause.

Two minutes. Ten trials. One reading.

Start the test