What HSK level is this text?

How to use the result

The bar shows how the words of your text spread across the HSK levels. The table under it answers a practical question: if you know every word up to a level, what share of this text do you know? Pick your level under “Mark words above” to see the words you are likely to miss, then copy them into your flashcards before you read.

For learners, a text where you know 95% of the words or more is good reading practice. For teachers, the list of words above the class level is the list to pre-teach.

Reading levels and HSK

The checker also names the reading level of the text in the same terms as the Dengdeng app: Beginner for roughly HSK 3.0 levels 1–2, Easy for roughly 1–3, Medium for 3–4, and Hard for level 5 and up. These are broad bands, not exam results.

What the checker cannot see

Word lists measure vocabulary. Grammar, long sentences, idioms and the topic also make a text harder. A text about a hobby you know is easier than its word counts suggest, and a text full of four-character idioms is harder. The checker counts each idiom as one word when it knows it.

For the sounds of the words, see the pinyin chart with audio. To look up a single word, use the Dengdeng dictionary, which also lists every word of each HSK level.

Questions

Which HSK lists does the checker use?

The HSK 3.0 word lists of 2021, with nine levels. Levels 7, 8 and 9 share one list and one exam, so the checker shows them together as HSK 7–9.

How does it decide the level of a text?

It splits the text into words and looks each word up in the HSK lists. A common word made of HSK words, such as 年轻人 (young people), takes the highest level of its parts. The level of the text is the lowest HSK level that covers 90% of its HSK words. Names, places and rare words are not in any list, so they are counted on their own and do not change the level.

Can I paste traditional characters?

Yes. Traditional spellings of HSK words are matched to their simplified form. The lists themselves are written in simplified Chinese.

Is my text sent anywhere?

No. The word lists load into your browser, and the text is split and checked there. Nothing is sent or saved.

Why does a word show a higher level than I expect?

The checker splits text by the longest words it knows. Sometimes a split differs from how a teacher would read the sentence, and a common character can sit inside a less common word. Short texts also swing more than long ones. Use the result as a guide, not a grade.

How much of a text should I know to read it comfortably?

Studies of reading in a second language put comfortable reading at about 95–98% of the words. At 90% you can follow a text with some lookups. Below that, most people stop enjoying it.

Data: HSK 3.0 word lists (GF 0025-2021) via the open hsk30 dataset (MIT); word frequencies from SUBTLEX-CH (Cai & Brysbaert, 2010, CC BY 4.0); pinyin and meanings from CC-CEDICT (CC BY-SA 4.0).