Performance Tests — Typing, Reaction Time & Speed
By Jayadeep · Last updated: August 2026 · How we test
Five benchmarks that measure you and your machine — typing speed, reflexes, aim, raw device performance, and battery state.
The tests on this page differ from the rest of the site in an important way: nothing here is looking for a fault. A microphone test has a right answer, and failing it means something is broken. A typing test has no right answer — it produces a number, and that number is only meaningful next to a baseline. So the useful question is not "did I pass" but "is this better or worse than last time, and how does it compare to people doing what I do?"
Two of these measure human performance. Typing speed is reported in words per minute, where a word is standardised as five characters including spaces — which is why WPM figures are comparable across tests with different passages. Accuracy matters more than raw speed for most real work, because corrections cost more time than they appear to; a fast typist with poor accuracy is often slower in practice than a moderate typist who rarely backspaces. Reaction time measures the interval between a visual cue appearing and your click registering, and it is worth knowing that the figure includes your display's latency and your mouse's polling interval as well as your nervous system. Comparisons are only fair on comparable hardware.
Aim training sits between the two, measuring accuracy and target acquisition rate rather than pure reflex. It rewards a different skill from reaction time — tracking and precision rather than raw speed — and improvement here comes from consistency of movement rather than from trying to move faster.
The remaining two measure the machine. The speed test runs a CPU workload and a graphics workload locally and reports operations per second and sustained frame rate, which is useful as a before-and-after when you suspect thermal throttling or want to compare two devices you actually own. The battery test reads level, charging state and estimated time remaining from the browser's Battery Status interface. That interface is deliberately limited for privacy reasons and is not exposed by every browser, so if it reports nothing, that is the browser declining to share rather than a fault with your battery.
Every performance test
Which test do I need?
Nothing here has a pass or fail, so pick by what you want to measure rather than by symptom. Run each more than once — single attempts are noisy.
| What you are seeing | Start here | Why this one |
|---|
| I want to know my typing speed | Typing Speed Test | Words per minute and accuracy over a timed passage. Take it more than once — a single run has a lot of noise in it. |
| I want to test my reflexes | Reaction Time Test | Measures the gap between a visual cue and your click. Your monitor and mouse contribute to the total, so compare like with like. |
| I want to practise aiming for games | Aim Trainer | Measures accuracy and targets per second, which rewards controlled tracking rather than raw speed. |
| Is my computer slow, or is it just me? | Device Speed Test | Runs CPU and graphics workloads on the device and reports a score. Useful as a before-and-after when you suspect throttling. |
| How healthy is my battery? | Battery Test | Reads level, charging status and estimated time remaining. Some browsers withhold this interface for privacy — no reading is not a fault. |
Troubleshooting guides
Context for interpreting your scores, plus practical advice on improving typing speed, reaction time and aim.
Frequently Asked Questions
What counts as a good typing speed?
Around 40 words per minute is typical for someone comfortable with a keyboard, and sustained speeds above 70 are genuinely fast for continuous prose. Context matters more than the headline number: professional transcription work expects considerably higher, while programming rewards accuracy and editing efficiency far more than raw WPM, because code is written in short bursts between thinking.
How is words per minute actually calculated?
A word is standardised as five characters including spaces, so total characters typed divided by five, divided by elapsed minutes. Using a fixed character count instead of real words is what makes scores comparable between tests that use different passages. Most tests, including this one, then adjust for errors — which is why accuracy affects your reported speed.
What is a normal reaction time?
Most people land somewhere in the region of 200 to 300 milliseconds for a simple visual cue. Your measured figure includes more than your reflexes: the time for the display to actually show the cue, and the delay before your mouse reports the click. A higher-refresh monitor and a higher-polling mouse will both improve the number without your reflexes changing at all.
Can I improve my reaction time?
Somewhat, and mostly through consistency rather than dramatic gains. Being properly warmed up, well rested and free of distraction moves the number more than most people expect, and practice reduces variance between attempts. The floor is set by physiology and there is no training that removes it.
Does the speed test compare my device to others?
No, and deliberately so. It runs a workload locally and reports what your device achieved. Publishing a cross-device leaderboard would require collecting and transmitting results, which is exactly what this site does not do. It is designed for comparing your own device over time or against another device in front of you.
Why does the battery test show nothing?
The Battery Status API is not implemented by every browser, and several removed it deliberately because battery characteristics can be used to help fingerprint a device. If your browser does not expose it, no reading is available — that is a privacy decision by the browser vendor, not a problem with your battery.
Other test categories
Audio tests · Screen tests · Controller tests · Input tests · Camera tests
Scores are calculated on your device and shown only to you. There is no leaderboard and no submission step, because publishing one would mean collecting results — which this site does not do. How we validate these tests.