Reaction Time Data: What 81,000 Real Tests Show
The median result on our plain visual test is about 325 ms. On the F1 start-light test it is about 397 ms. Audio is faster than both at 307 ms, and the color test is slowest at 470 ms.
These numbers come from 81,106 completed tests on this site between 18 July and 20 October 2026. Each result is the average of five attempts. Nothing here is a lab study, and we say so in the method note. If you want to add your own result to the pile, take the reaction time test.
Published 21 October 2026. Data from 18 July to 20 October 2026.
How fast is each type of test?
Audio has the lowest median, 307 ms, and the visual test follows at 325 ms. Both are simple tasks: one signal, one click. The F1 test is also one signal, but you wait through five red lights first. The color test asks you to pick the right color, which is a choice task and takes longer.
| Test | Results | Median | Middle half | Under 250 ms | Under 200 ms |
|---|---|---|---|---|---|
| Visual | 47,888 | 325 ms | 274 to 407 ms | 12.5% | 0.9% |
| F1 start lights | 28,080 | 397 ms | 329 to 550 ms | 5.2% | 1.4% |
| Audio | 2,993 | 307 ms | 258 to 381 ms | 20.1% | 1.5% |
| Color | 2,145 | 470 ms | 408 to 539 ms | 1.8% | 0.7% |
Audio and color are small samples (about 3,000 and 2,100 results), so treat their medians as a good estimate, not a precise one. We left out the aim test, because it times how long you take to hit targets, which is not a reaction time.
How does the F1 start-light test compare with the plain visual test?
F1 players are slower on average and far more spread out. The median is 397 ms against 325 ms, a gap of about 70 ms. The visual results bunch up: 47% land between 250 and 349 ms. The F1 results spread out, and 17% are 700 ms or slower, against 6% on the visual test.
Which share beats 200 ms?
Very few, on either test. An average under 200 ms is rare because it means all five attempts were fast. On the F1 test 1.4% of results beat 200 ms. On the visual test 0.9% did. The F1 test has a slightly larger share of very fast results and a much larger share of slow ones.
We cannot tell why F1 results run slower. Our guess is the waiting: the lights take a few seconds to go out, and a lot of people click on instinct or lose focus. We did not test that. If you want to see how you do, try the F1 reaction time test and compare it with the visual test.
Method note
- What is counted. One result is the average of five attempts in one test, sent when the fifth attempt ends.
- Who is in it. People who found the site and chose to take a test. They are not a random sample of the population, and we do not know their age or sex. Visitors who send Do Not Track or Global Privacy Control are not counted.
- Rounding. We store each average in 25 ms steps, so medians and percentiles are estimates inside a step, accurate to about 12 ms. Averages under 100 ms and over 1,500 ms are stored as 100 and 1,500.
- Device delay. A browser test adds delay from the screen, the mouse or touch screen and the browser. Display and input delay can add 10 to 30 ms, which is part of why our visual median sits above the 250 ms often quoted from lab studies, listed on the average reaction time page.
- Privacy. The data is anonymous counts. We store the test type and the rounded average, with no name or visitor ID. Groups with fewer than five results are merged before we see them, so nothing small is published.
- What is missing. We do not yet publish how results change from one test to the next within a visit. We only started recording the run number on 19 October 2026, so the sample is still too small. We will add it when it is large enough.
You are welcome to quote these numbers. Please link to this page, and say the data comes from reaction-time-test.io. Questions go to michael@1000ventures.io.
