Home / Words and characters for fluency
Reference · computed from the word lists
How many words and characters do you need for daily life, media, and fluency?
Quick answer
How many words and characters do you need to get by in Chinese?
Framed by HSK level: everyday survival is about HSK 3.0 Level 3, roughly 2,209 words and 899 characters, recognising near 66.8% of everyday speech. Following most media sits around Level 6 (5,363 words, ~73.9%), and wide, near-native reading is the full 7 to 9 band (10,969 words, 2,970 characters).
- Free, always. No account, no email, no paywall.
- Computed, not asserted. Words, characters, and coverage all come from the shipped HSK word lists, so nothing drifts from the words the site serves.
- Sourced. Spoken coverage is measured against a published film and TV frequency list. Last verified 2026-07-24.
What each HSK level actually buys you
interactive · pick a levelWords + characters + spoken coverage, by level
Share of running words in everyday speech you would recognise
Hold basic conversations, manage travel, work, and study small talk.
Words and characters are cumulative: each level adds to every level below it. Spoken coverage is the cumulative share of Chinese running words in the OpenSubtitles 2018 corpus, matched against 99.5% of the HSK 3.0 list. Recognising most words is not the same as understanding every sentence. See characters per level and the full word tables.
Every HSK 3.0 level: words, characters, coverage
Daily life, media, fluency: where the lines fall
Get by day to day
HSK 3
Everyday survival: shops, travel, small talk, and the bones of a conversation.
Follow most media
HSK 6
News, films, and general articles become mostly followable with a little help.
Read widely, near-native
HSK 7-9
Literature, opinion, and specialist prose across the full advanced band.
The three lines are not sharp thresholds, they are where the curve does most of its work. The first HSK 3.0 level already recognises about 50.4% of everyday speech, because a small set of very common words does most of the talking. After that the coverage climb flattens: each new level adds thousands of rarer words for a few more percentage points. That is why "daily life" arrives early on the ladder and "read anything" never quite arrives, even at the top.
How many characters is that, really?
Characters climb more slowly than words, because one new word often reuses characters you already know. The HSK 3.0 standard sets an official character syllabus of 300 recognition characters per level, reaching 1,800 through Level 6 and 3,000 across the full 7 to 9 band. This site does not assert those round targets: it counts the distinct characters that actually appear in the words it serves. The two line up closely.
Computed characters vs the official HSK 3.0 syllabus
| Level | This site (computed) | Official 3.0 syllabus | Difference |
|---|---|---|---|
| HSK 1 | 300 | 300 | +0 |
| HSK 2 | 598 | 600 | -2 |
| HSK 3 | 899 | 900 | -1 |
| HSK 4 | 1,199 | 1,200 | -1 |
| HSK 5 | 1,499 | 1,500 | -1 |
| HSK 6 | 1,799 | 1,800 | -1 |
| HSK 7-9 | 2,970 | 3,000 | -30 |
Reconciling the gap honestly. Through Level 6 the computed counts sit within a handful of the official quota (598 against 600, 899 against 900, and so on): the tiny shortfall is official syllabus characters that never happen to appear in any word of the shipped list. Across the full scheme the site computes 2,970 distinct characters, just under the 3,000-character target for the same reason. A third figure is often quoted, a recognition-character progression of about 246 characters at Level 1 rising to roughly 3,088 by the advanced band. That set counts the distinct characters actually used across the official vocabulary rather than the round syllabus: 246 is below 300 because the official Level 1 word list does not exhaust all 300 Level 1 characters, while ~3,088 is above 3,000 because a few words reach for characters beyond the syllabus. This site's list is a community aggregate that differs slightly again, front-loading a full 300 distinct characters into its larger Level 1 word set. Three vantage points, all within a few percent, none pretending to be the others.
Coverage is not comprehension
Recognising three-quarters of the words spoken is not the same as understanding three-quarters of what is said. The unknown slice tends to hold exactly the low-frequency words that carry the new information in a sentence, so the last stretch toward comfortable, unaided understanding is far harder than the coverage percentage alone suggests. Treat these figures as a map of the climb, not a promise of fluency at any given rung.
Want to know which rung you are actually on before you plan around any of these numbers? Take the five-minute placement test, browse the word tables by level, or see the characters and radicals reference and the words-per-level guide.