Typing Tests
Twenty-one typing tests grouped by what each one varies: the length of the sample, what the text is made of, or what a mistake costs you.
If you only want a number, take the one-minute test and stop reading. Sixty seconds is the window nearly every typing speed you have ever been quoted was measured over, so it is the figure that compares to other people's figures. Everything else on this page exists for a different question: not how fast you type, but what your speed is made of.
There are only three things a test can vary, and every test here varies exactly one of them. It can change the length of the sample, which changes how much a single fumbled word matters and whether fatigue gets a say. It can change what the text is made of, which changes which fingers are doing the work. Or it can change the rule about mistakes, which leaves the typing identical and changes what the score means. Read in that light, a run of tests is a set of experiments on yourself rather than a leaderboard.
Why the same typing scores differently at different lengths
Words per minute is a rate, and a rate has a denominator. Over fifteen seconds you might type fifteen or twenty words in total, so one word lost to a fumble is a large fraction of everything you produced and the arithmetic magnifies it — a single extra correct word can move a fifteen-second result by more than it could move a one-minute result four times over. That is not noise in the sense of a broken measurement; it is what a small sample of anything looks like.
Going the other way introduces something a short test cannot detect at all. Past a couple of minutes your accuracy usually starts drifting before your speed does, and by ten or twenty minutes the limiting factors are attention and hand recovery rather than technique. That is why the twenty-minute test is not simply a longer version of the short ones: the useful figure it produces is the gap between your best and worst minute, which no short test has enough minutes to give you. If you want the arithmetic behind all of this written out, the guide on how WPM is calculated has the formulas.
What the content modes are really testing
The content-mode tests exist because a single WPM figure hides an enormous amount. Almost everyone is dramatically slower on digits than on letters, because the number row is learned last and offers no word shape to anticipate. Almost everyone is faster on the hundred commonest words than on ordinary prose, because vocabulary has been removed as a variable. Neither of those numbers is your real speed; the difference between them is the interesting part.
The pairing worth running deliberately is lowercase-only against your ordinary result. One removes every modifier — no shift, no punctuation — so the gap between the two prices what modifiers cost you personally, rather than telling you in general that they cost something. If that gap is large, the shifted top row and sentence punctuation drills are where it gets closed.
Two tests that change the rule instead of the text
The no-backspace test disables correction entirely, so an error is permanent and the score is what you actually produced rather than what you managed to repair. Most people find their net speed barely moves and their confidence collapses, which is itself the finding: it reveals how much of an ordinary session was spent fixing rather than writing.
Accuracy mode applies the opposite rule — you cannot advance past a mistyped character until it is corrected — so the passage cannot be finished with errors in it and time becomes the only variable. Error rate stops being a percentage and becomes seconds on the clock, which is a considerably harder thing to argue with. The reasoning behind that trade is set out in accuracy versus speed.
Reading a result honestly
Three habits make a typing result mean more than it usually does, and all three are free. The first is to take one attempt rather than several. Running a test five times and keeping the best is selection rather than measurement — it produces the highest number your variance permits, which is not a figure you can compare against anything later, including your own result next month.
The second is to warm up before the attempt that counts. Your first minute of typing is reliably slower than your third, for reasons that have nothing to do with skill: your hands are still settling, your reading has not yet run ahead of your fingers, and your attention is still arriving. On a short test that whole effect sits inside the measurement. Two minutes of untimed typing beforehand removes it, and warming up and taking breaks covers what else changes across a session.
The third is to keep the conditions constant. A result from a fresh morning and one from the end of a long afternoon are not two points on the same line, and neither is wrong — they answer different questions. Pick one and stay with it, because the trend is worth far more than any individual figure.
Why your score here may not match another site
Two sites can watch the same minute of typing and report different numbers without either being wrong. A word is five characters almost everywhere, and that is roughly where the agreement ends.
The larger disagreement is about errors. Some sites count every incorrect keystroke, whether or not you fixed it; others count only what is still wrong when the clock stops. A typist who corrects everything scores very well under the second rule and mediocre under the first, from identical typing. Then there is the error penalty itself, which is applied per minute on some sites and simply by not counting wrong characters on others — and the timer, which starts on page load in some places and on your first keystroke in others.
This site counts uncorrected errors and starts the clock on your first keystroke, so reading the opening line costs you nothing and corrections cost you time rather than accuracy. The full arithmetic, with a worked example producing three defensible numbers from one session, is in how WPM is calculated.
Which test to take, by what you are asking
If you want a number to quote, take the one-minute test. It is the window nearly every typing speed anyone has ever mentioned to you was measured over, which makes it the only figure that compares to other people's figures at all.
If you want to know whether your typing holds up, take something long. The five-minute test is the shortest one where sustained accuracy rather than a good opening decides the result, and the twenty-minute test is where attention and hand recovery start to dominate technique entirely.
If you are preparing for an assessment, match the format rather than the number — length, content and error rule all move a result, and typing tests in hiring lists what to ask an employer beforehand. And if you suspect a specific weakness, pick the mode that isolates it: digits, punctuation, capitals or code each remove everything else from the picture, which is what makes them diagnostic rather than merely harder.
By length
Same kind of text, different sampling window. A short window amplifies variance; a long one introduces fatigue and drift that a short one cannot see at all.
By what you type
The character mix changes which fingers do the work, and whether the five-character word is even a fair unit for what you produced.
By the rule
Same text, same clock, different definition of what a mistake costs — which changes what the score means rather than how hard the typing is.
After the number
A test tells you where you are; it does not move you. Every result here records which individual keys you missed, and that per-key map is the thing worth acting on rather than the headline figure. If it points somewhere specific, the drills are organised by exactly the categories it will point at — a finger, a letter pair, a punctuation mark, a symbol set. If it points everywhere at once, or if the keys it flags are simply the ones you typed most, run the weak-key diagnostic, whose text touches every key at roughly even frequency so the map it produces is not just a picture of English letter frequency.
And if you are not yet typing without looking down, no test on this page is the right next step. The practice path is, and it will make every number here better than repeating the tests would.