How fast do people type in Japanese?

The best measured figures put Japanese typing at about 17 words per minute on a cellphone and 23 words per minute on a desktop PC ^{1 2}. Those numbers come from one small, dated study of copying, not from a survey of typical typists. For a benchmark of skilled typing, a certification test sets pass levels, and the 2nd level requires about 60 characters per minute ³.

For the full picture of how Japanese speech and typing compare, see the guide to Japanese, Chinese and Korean speech to text, and for English speeds, see the typing speed overview.

The best measured Japanese speeds are 17 to 23 words per minute

In the study, 24 Japanese university students copied passages for two minutes. They typed Japanese at 23 words per minute on desktop PCs and 17 on cellphones ^{2 1}. The same students also typed on miniature thumb keyboards and onscreen keyboards, which were slower. The study is small and the devices are dated, so use the figures as a rough guide to copying on those devices, not as a current average.

A copying test is not the same as writing your own text. Choosing words, correcting your own sentences and thinking about what to say all slow a typist down.

Japanese speed is counted in characters, so it is not directly comparable to English

Japanese has no spaces between words, so there is no clean word count. Typing tests therefore count characters, and the usual English measure of words per minute does not transfer. For example, the 51.56 words per minute average in a large English keyboard study used a word defined as five characters ^{4 5}. That figure describes English typing, and it does not describe Japanese typing.

Japanese typing also adds steps that English does not have. You enter kana, then choose the kanji or kana form from a list, because one sound can be written several ways. The choices take time, and the study data do not measure how much of the total time they account for. We did not find a published keystroke count for Japanese kanji conversion in the sources we opened, so the size of that cost is not measured here.

A certification test sets the pass marks in characters per minute

The Japan Information Processing Certification Association runs a 10-minute Japanese text input test, and its pass levels are set as net characters typed in that time ⁶. The 2nd level needs 600 net characters, which is about 60 characters per minute ³. The 1st level needs 1,000 characters in the same 10 minutes ⁷.

Two rules matter when you read those thresholds:

• Each mistake costs one character. The net count is the total typed minus the mistakes ⁸. A fast typist who makes many errors can score below a slower, accurate one.
• The test is copying. Candidates type the question text as written, not their own words ⁹. Copying removes the time needed to choose words and kanji, so a passing score on this test does not show how fast you compose.

The thresholds are not averages. They show what the association requires for each level, not how many people reach it.

It depends on the device, the input method and the task

Speed depends on the device and the task. In the study above, the same students were faster on desktop PCs than on cellphones for the same two-minute copying task ^{2 1}.

Speech is a separate measure, and this question covers typing only. If long Japanese text is the bottleneck, dictation is the other route to written Japanese, and the guide covers it.

Korean is a separate case. We searched for measured Korean typing speeds in words per minute and did not find a published study of hangul keyboards in the sources we opened. The practice material we found gives target speeds without any data behind them, so we leave those targets out.

What to do: measure your own speed, then choose the right tool

1. Measure your speed the way the certification does. Type a passage you know for 10 minutes, count the characters you typed, and subtract one for each mistake. Compare the result with the 600 and 1,000 thresholds above, and remember that the test measures copying, not composing.
2. Compare like with like. Measure copying speed against copying speed, and composing speed against composing speed, on the same device.
3. If writing long Japanese text is the bottleneck, try dictation. Dictation can save the time spent choosing candidates, though you still check homophones. The guide to Japanese, Chinese and Korean speech to text explains where it works well and where homophones cause errors, and the best speech to text for Japanese compares the options.
4. Check the output. Dictation errors in Japanese are often homophones, so read the result before you send it, and count any corrections in your total time.

For the other side of the comparison, see whether you can type faster than you talk, how the average typing speed varies by age and how words per minute is calculated. The terms used here are explained in words per minute, keystrokes per character and character error rate.