How fast do Korean learners type?

Speed by mode and TOPIK level

Each bar spans the 25th to the 75th percentile, so half of all sessions fall inside it. The black tick is the median.

Most practice sessions are spent on the keyboard-row drills, which carry no TOPIK level. Only 175 sessions (practice and test) were mostly TOPIK word lists, and those run faster, probably because learners move on to whole words once they have found the keys.

The typing test median of 25 WPM comes from just 40 sessions, and the people who choose to take a timed test are probably the more confident typists. Treat it as the speed of learners who test themselves.

Accuracy barely moves between groups: every median on this page sits between 89% and 95%. Speed is where learners differ.

Speed by country

Countries with at least 20 sessions, ranked by median WPM. Most groups are small, so a difference of one or two WPM between them means little.

South Korea has the second largest group and a wide spread (11.6 to 23.3 WPM). It likely mixes learners living in Korea with Korean speakers trying the site, and we cannot tell them apart. Country comes from the device time zone, and a device we cannot place counts as the United States, so the US row also holds some visitors from elsewhere.

WPM, 타수 and syllables per minute

Three units describe Korean typing speed, and they measure different things:

• WPM counts keystrokes, five to a word, per minute. It is the unit English typing tests use and the one this site shows by default.
• 타수 (tasu) is keystrokes per minute, the number Korean typing programs show. On our scale 1 WPM = 5 타. Programs differ on how they count Shift and Backspace, so expect a small gap between tools.
• Syllables per minute counts finished Hangul blocks: 학교 is two. Korean exams count characters, so this is the unit that maps to an essay.

The 1,013 words in our TOPIK 1 and 2 lists average 2.47 keystrokes per syllable: 가 takes two keys, 한 three, 왔 four. Add a space after each word and running text comes to about 2.89 keystrokes per syllable.

Shaded rows are the practice median (12 WPM) and the typing test median (25 WPM). The typing test uses text with spaces and reports syllables per minute directly.

Rough speed levels for learners

These bands are our own reading of the data above, not an official scale.

For scale: a study of 168,000 volunteers typing English found an average of 52 WPM (Dhakal et al., CHI 2018). Both use five keystrokes to a word, but a Korean keystroke carries a different share of a word than an English one, so the comparison is loose.

What the TOPIK IBT essay needs

The TOPIK II IBT essay asks for 400 to 500 characters, spaces included, inside a 50-minute writing section (official IBT demo). A full 500-character answer holds about 390 syllables and 110 spaces, close to 1,075 keystrokes. That is how long the typing alone takes, before any planning or rereading:

At the site-wide median, typing the essay uses more than a third of the section. At 17 WPM or more it takes under 13 minutes, which leaves time for the short answers and a reread. TOPIK publishes no speed requirement; the full reasoning and a two-week plan are in the TOPIK typing guide.

How we collected the data

• Each practice visit or completed typing test sends one summary row: mode, WPM, accuracy, the TOPIK level of the words, country and length. No name, account, IP address or device ID is stored (privacy policy).
• WPM counts correct keystrokes only, five to a word, per minute of active typing. A gap of more than 3 seconds is treated as a break and not counted. Sessions under 15 seconds are dropped.
• The rows cover late October to November 4, 2026. A person who visits five times counts five times.
• The numbers come from an aggregate-only database function: it returns percentiles, never single rows, and leaves out any group with fewer than 20 sessions. That also means we could not hand-check individual rows. Our automated tests never send rows, but a handful of sessions from our own development may be included.
• This is everyone who used the site, not a controlled sample. Beginners on keyboard drills pull the overall median down; confident typists are over-represented in the typing test.

We plan to refresh these numbers as more sessions come in. To cite them, link to this page and give the "data as of" date above. Questions: hi@typehangeul.com.