Average typing speed, and what counts as good

Everyone quotes about 40 words a minute. Almost nobody says which of the four formulas that number came out of, and the gap between them is bigger than the gap between a slow typist and a fast one.

The short answer

The commonly quoted average is around 40 words a minute, and a good speed is around 60 at 97% accuracy or better. Both figures are conventions rather than measurements, and both are published without the one piece of information that would make them mean anything: which definition of a word per minute was used.

That is not a pedantic complaint. We ran one identical, exactly-timed piece of typing through all four published formulas. At 95.1% accuracy it scored anywhere from 70.6 to 93.7 words a minute, a spread of 24.7%. The measurements are below, with the method beside them.

Why two tests give you two different numbers

Take the same test twice on two sites and you will usually get two answers. Three things differ between them, and only the third is about you.

1. The formula

   Four are in common use. All four call five characters a word, and that is the only thing they agree on. What they disagree about is how to price a mistake.

2. Where the clock starts

   Some tests start counting on your first keystroke. Others start when the page finishes loading, so the seconds you spent reading the instructions are in your denominator.

3. The text

   A list of common three-letter words with no capitals in it is far quicker to type than prose with commas and shift keys in it. The same typist can be 15 words a minute apart on the two without doing anything differently.

The four formulas, side by side

What we measured, and how

Run on 30 August 2026 in Chrome 151 on macOS, using the floi typing test in its Custom mode so the prompt text was identical every time.

Rather than type by hand, we drove the tool with a scripted keystream, waiting on an absolute clock between keystrokes so the rhythm could not drift. The keystrokes went into the tool's real input, so its own counters produced the totals. In the first table the interval is a fixed 130 milliseconds and the only thing that changes between runs is how many characters were typed wrongly. No error was ever corrected.

The same run, priced four ways

The passage was 203 characters. Every run took 26.0 seconds and 203 keystrokes. Nothing varied except how many of those keystrokes were wrong.

Three things fall out of that table, and the first is the one nobody mentions.

• At 100% accuracy the three formulas are identical. All the disagreement between typing tests is a disagreement about mistakes. Type a perfect run and every site on the internet will hand you the same number.
• Gross never moves. Ten wrong characters cost exactly nothing, because gross speed measures how fast your fingers went and not what came out. A site reporting a suspiciously flattering number is probably reporting this one.
• Adjusted punishes an error about five times harder than net. Across those ten errors, net lost 0.46 words a minute each and adjusted lost 2.31 each. At 95.1% accuracy, which most people would call a decent run, the two are 18.5 words a minute apart on identical keystrokes. The adjusted figure sits 20.8% below the net one, a wider gap than the one between an average typist and a good one.

What one typo costs, by test length

The adjusted penalty is a rate rather than a fixed amount, so the shorter the test the more a single mistake hurts. Three passages, exactly one uncorrected error in each. The three ran at different speeds, which is why the gross column moves, and it makes no difference to the point: the penalty is one word divided by the minutes, so only the length of the run decides it.

Eight times the length, and the same single typo costs roughly nine times less. This is why a fifteen-second test is entertainment and a five-minute test is an assessment. It is also why a bad fifteen-second score is worth ignoring: you did not get slower, you got unlucky about where the mistake landed.

The clock is the third disagreement

Every number above divides by the time from the first keystroke to the last. Some tests divide by the time since the page appeared instead, which quietly bills you for reading the instructions.

On the clean 26.0-second run, 203 correct characters scored 93.7 words a minute. Add a three-second pause before the first keystroke and the identical typing scores 84.0, because the denominator went from 26 seconds to 29. That is 9.7 words a minute, or 10.4% of the score, for three seconds of reading the page.

If a site gives you a number well below what you usually get, click into the field, wait, and see whether the timer has already started. That one check explains a lot of surprising results.

So what is a good typing speed?

With all of that in mind, here are the bands as they are actually used in hiring and training. They are conventions and not measurements, and they assume a net figure at 97% accuracy or better.

Average by age and by job, and why to distrust both

There is no census of typing speed. The by-age and by-profession tables that circulate all trace back to a small number of studies and vendor datasets, none of which publish their formula, their text or their accuracy threshold.

Given the tables above, a by-age figure with no formula attached could be describing typing 23% faster or slower than the reader assumes. We are not going to add another one. What is worth knowing is the shape rather than the numbers:

• Speed tracks practice, not age. The strongest predictor of how fast somebody types is how many hours a day they type, which is why students and office workers in the same age band are far apart.
• The big jump is technique, once. Moving from looking at the keyboard to not looking at it is worth more than every other improvement combined, and it happens over weeks rather than years.
• Job requirements are about accuracy as much as speed. A data entry role asking for 60 words a minute usually wants 98% with it, and under the adjusted formula that second number is the harder one.

Which number a job application means

If a form asks for your typing speed, it almost certainly means a net or adjusted figure from a timed test of at least a minute, at a stated accuracy. Quote the number from a longer test rather than your best fifteen-second sprint, and say the accuracy beside it. "68 wpm at 98%" is a claim somebody can check. "Around 80" is not, and it is the kind of thing that gets checked at the interview.

If the employer runs their own test, expect it to be five minutes and expect it to use the adjusted formula. Run a five-minute test yourself first: it is a genuinely different experience from thirty seconds, and it is where most people find out that their real constraint is concentration.

How to actually get faster

Four things, roughly in order of what they are worth.

1. Stop looking at your hands

   Everything else is a rounding error next to this. It is uncomfortable for about a week and your speed will drop while you do it. That drop is the price of the whole gain.

2. Slow down until you are accurate, then speed up

   Typing at 97% and building speed is faster than typing at 90% and trying to clean it up, because the mistakes are habits and habits get faster too.

3. Find the keys you hunt for, and drill those

   Most people lose more time to four or five keys than to everything else combined, and almost nobody can name which ones. The Keys tab on our typing test shows the milliseconds it took you to reach each key, so the list stops being a guess.

4. Practise on the text you actually type

   A word list has no capitals, no commas and no brackets in it. If your day is email, drill the paragraph mode. If it is code, drill the brackets. Paste your own text in if neither fits.

The measurements on this page were run on 30 August 2026, in Chrome 151 on macOS, with a scripted keystream at a fixed 130 millisecond interval. Every figure comes from the tool's own counters on that run. Browsers and tools change, so if you repeat it and get something different, trust what you see.