Average WPM and Benchmarks — How Many Characters per Minute? Japanese Typing Speed in Official Surveys and Real Data

WPM stands for Words Per Minute. The 'word' here is not a Japanese word. It is an agreement to count five characters (five keystrokes), including spaces and punctuation, as one word, a convention from English text entry.

This is a widely used convention, not an international standard. I could not find a standard that defines it. For Japanese romaji typing, the usual practice is to divide the number of keystrokes by 5.

Here is how Typing Musou counts. It helps when you compare with numbers from other sites.

• It counts romaji key presses

  Not the number of kana characters. Typing し as 'shi' is 3 keystrokes; typing it as 'si' is 2. The same text gives a different WPM depending on the spelling you choose.

• Mistyped keys are not counted

  Only correctly typed keystrokes go into the numerator. Time spent stopped after a mistake stays in the denominator. Some methods count all keystrokes including mistakes; the WPM in this article is the net figure without mistakes.

• Other sites' 'WPM' may be defined differently

  For example, e-typing's official help describes WPM as 'the number of characters (alphanumeric keys) entered per minute' (checked on 2026-10-18). It does not mention dividing by 5. If taken literally, that is the keystroke count itself, about five times the WPM in this article. Before comparing numbers across sites, check the definition on each result screen or help page.

When someone asks 'how many characters per minute', it depends on what counts as a 'character'. Four units come up often.

A fixed ratio between kana and keystrokes is impossible because each kana takes a different number of keystrokes. あ is one keystroke (a). し is 2 keystrokes as 'si' and 3 as 'shi'. きょ is 3 keystrokes (kyo) for two kana.

Counting displayed characters widens the gap. For example, 学校 is 2 characters on screen, but typing がっこう takes 6 keystrokes. Text with many kanji tends to need more keystrokes per displayed character.

So 'WPM 40 equals X characters per minute' cannot be stated. What can be said is that WPM 40 is 200 keystrokes per minute. As a guide, WPM 20 is 100 keystrokes, 30 is 150, 50 is 250 and 60 is 300. If you apply the range of the five examples above (1.0–1.7 keystrokes per kana), 200 keystrokes correspond to roughly 120–200 kana. That range only comes from these examples and does not hold for every text.

This is the distribution from Typing Musou records (Aug 3 to Oct 18, 2026; 4,942 players, 48,215 plays). It covers people who played the 60-second Speed Trial. Everyone plays under the same conditions: the samurai, no star upgrades, no ultimate skills, scoring by damage in 60 seconds. I took each player's median WPM and looked at the distribution of people.

This article uses the value of the person in the middle (the median). The mean tends to be pulled up by fast people. The MEXT survey also shows a mean higher than its median.

• These are people who played a game

  This is not an average of all people in Japan. We do not know age or job, so it cannot be called an average for students or office workers. Records of players who turned off research use, players who left, and students who joined a school class are excluded.

• The most struggling players' plays are not included

  Plays with accuracy of 70% or below were not recorded. The whole group of people who tried the game may extend further toward the slower side than this distribution.

• People who keep playing tend to be faster

  Limiting to the 1,929 players with five or more plays, the median is 44 (22 at the bottom 10%, 76 at the top 10%). This does not show that playing makes you faster; fast players may also be the ones who keep playing.

The middle player's first play was 38 and personal best was 43, so even the same person varies from play to play. The median accuracy per player was 92.5%.

What the data supports is this much: in the Typing Musou Speed Trial, the middle player is at about 40 WPM and the top 10% are at about 70 WPM or more.

The band names are this article's own, assigned by where each band falls in the distribution above. They are not an industry or official standard. They show a position among people who played Typing Musou, not among all people in Japan.

How many WPM is enough depends on your goal. For speed at work, see How Fast Do You Need to Type for Work?; official targets are in the next chapter. If your speed has stalled, read Why Your Typing Isn't Getting Faster.

It is not true that no official typing survey exists. MEXT's Information Literacy Survey measures keyboard text entry by students. The fiscal 2021 survey (carried out January–February 2022, published March 2023) covered about 14,000 students in Grade 5, Grade 8 and Grade 11.

The task was to type a 285-character text, mixing kanji, hiragana, katakana and letters, in three minutes. Students were told they could skip characters they did not know how to enter. Results are reported as characters entered per minute.

In the same survey, 2.3% of elementary, 6.4% of junior high and 13.9% of high school students could enter one A4 line (about 40 characters) per minute. Those under 5 characters per minute were 17.6%, 3.8% and 1.1%.

These figures are characters per minute: characters of the task text (mixed kanji and kana) entered per minute, not keystrokes. They cannot be converted to WPM (keystrokes ÷ 5), and MEXT itself does not convert them. If you see MEXT's numbers turned into WPM, check what assumptions were used. For grade-by-grade views, see When Should Kids Start Typing? and Typing Speed of Middle & High School Students.

I could not find a national statistic of Japanese typing speed for adults in the materials I checked from the internal affairs ministry and MEXT. MEXT's survey covers only Grade 5, Grade 8 and Grade 11 students.

Typing Musou does not ask for age or job either, so a distribution for office workers alone cannot be produced. The distribution in chapter 3 is for everyone who played.

When you see a figure like 'the average office worker types X WPM', check whether the population, sample size and WPM definition are given. For Japanese typing speed, the only sources I could trace for this article are MEXT's student survey and Typing Musou's own records.

I also found no nationally unified standard for the speed needed at work. The Ministry of Health, Labour and Welfare's guideline for information-device work, as far as I read it, has no typing-speed standard (it does set rules for work time, breaks and posture). For ways to think about each job and how to set a target from your own typing volume, see How Fast Do You Need to Type for Work?. For certification standards, see Typing Tests in Japan.

A large English-language study was published in 2018 (Dhakal et al., CHI 2018). About 168,000 people took part online, voluntarily, transcribing 15 English sentences. The mean was 51.56 WPM (SD 20.2). The slowest 10% were under about 26 WPM, the fastest 10% above about 78 WPM, and the fastest typists reached 120 WPM or more.

However, the participants were self-selected, 68% were from the US, and the mean age was 24.5. They are not representative of the general population.

Please do not compare this number with a Japanese romaji WPM. The task, the people and the key sequences are all different. Even when the formula is the same, a bigger or smaller number does not simply mean that Japanese typists are slower or faster. I found no study that compares Japanese and English WPM. For English practice, see Tips for English Typing.

To measure, you can use the Typing Musou Speed Trial. It is a mode where you score by damage dealt to a sandbag in 60 seconds. It is not a screen made only for measuring WPM, but the result screen shows WPM and accuracy. The only number you compete on is damage, and it also shows the firepower you missed because of mistakes.

• Measure several times under the same conditions

  Take a few measurements with the same mode, keyboard and time of day, and the number settles. The first play is affected by novelty and nerves; in our records, personal bests were higher than first plays (38 and 43 for the middle player). Looking at the median of a few plays is my advice as the developer, not a method confirmed by research.

• Look at WPM and accuracy together

  WPM counts only correctly typed keystrokes. More mistakes lower the number, because time stopped after a mistake also stays in the denominator.

• Spelling choice is also a condition

  Typing し as shi or si changes the keystroke count. When comparing, measure with the same spelling.

• Do not line up numbers from different sites

  WPM is defined differently by site (chapter 1). To track progress, collect your numbers on one site.

• Q. Is 50 WPM fast?

  Among people who played the Typing Musou Speed Trial, about 3 in 10 (31%) were at 50 WPM or more. 50 WPM is 250 keystrokes per minute, about 0.24 seconds per keystroke. Where it stands among all people in Japan cannot be said, because there is no public statistic.

• Q. Can I trust figures like 'the average Japanese person types X WPM'?

  Judge by the source. A number without the population, sample size and WPM definition cannot be used as is. For Japanese typing speed, the only sources I could verify for this article are MEXT's student survey (characters per minute) and the records of people who played Typing Musou.

• Q. What is the average WPM for elementary or junior high students?

  The official figures are in characters per minute, not WPM: 15.8 for Grade 5, 23.0 for Grade 8 and 28.4 for Grade 11. Typing Musou also does not use the records of students who joined a school class in its tallies, so no grade-by-grade WPM is given.

• Q. Do mistakes lower WPM?

  Yes. WPM counts only correctly typed keystrokes, and time stopped after a mistake stays in the denominator. For ways to reduce mistakes, see How to Reduce Typing Mistakes.

• Q. How rare is 100 WPM?

  In the Typing Musou Speed Trial, about 1.4% of players had a median of 100 WPM or more, and the top 1% position was about 103 WPM. That is the position among people who played the game. It cannot be compared with the English study's figures.

WPM is correctly typed keystrokes ÷ 5 ÷ minutes. Counting five keystrokes as one word is a convention, not a standard. 'Characters per minute' depends on whether you count kana, displayed characters or keystrokes, so no fixed conversion exists. Definitions also differ by site, so check them before comparing numbers.

I could not find a national average for Japanese typing that includes adults. The official figures are MEXT's student survey (Grade 5: 15.8, Grade 8: 23.0, Grade 11: 28.4 chars/min). In records of people who played Typing Musou, the middle is about 39 WPM and the top 10% are at about 71 WPM or more. That is a record of game players, not an average of all people in Japan.

First, measure your WPM once in the Speed Trial. You will see where you sit in the distribution.

I checked the official figures and other studies against the documents below (checked October 18, 2026). The MEXT and MHLW documents are in Japanese. Page numbers are PDF page numbers. The Typing Musou records are my own tally (October 18, 2026; only players who allowed research use).

1. MEXT, Survey on Information Literacy: Results (March 2023) — p.56–57 (Figures 2-4-1 and 2-4-2: mean, median and distribution of characters entered per minute; the 285-character, 3-minute task; share who could enter one A4 line)
2. MEXT, Special Committee on Curriculum Planning, Document 1-1 (May 12, 2025) — p.15 (current value and target for average characters entered per minute: 40 for elementary and 60 for junior high by 2026 (Reiwa 8; the source writes "R8"); the target comes from the Education DX KPI, April 22, 2024, Digital Administrative and Fiscal Reform Council)
3. Ministry of Health, Labour and Welfare, Guideline on Occupational Health Management for Information-Device Work (commentary) — to confirm that it sets no typing-speed standard (it does set rules for work time, breaks and posture)
4. e-typing help: “What are WPM and EPM?” — how it explains WPM (“the number of characters (alphanumeric keys) entered per minute”)
5. Dhakal, Feit, Kristensson, Oulasvirta. Observations on Typing from 136 Million Keystrokes. CHI 2018 — about 168,000 people transcribing 15 English sentences; mean 51.56 WPM (SD 20.2); slowest 10% under about 26 WPM, fastest 10% above about 78 WPM; participant demographics