Esta página también está disponible en español. Leer en español →
Cette page est aussi disponible en français. Lire en français →
Diese Seite gibt es auch auf Deutsch. Auf Deutsch lesen →
本页面也有中文版。 阅读中文版 →
The HSK 3.0 syllabus, as files
Everything this site publishes, in formats you can take away: 11,000 words, 3,088 reading characters and 428 grammar points. Per level, or all of it in one file.
Every row carries the word, pinyin with tone marks, the tone number of each syllable, the English meaning, an example sentence in Chinese, pinyin and English, and the address of a recording of the word. That last column is the one no other published list has: 13,390 clips, free to play.
Two formats. The CSV opens in Excel, Numbers and Sheets, and in anything else that reads a spreadsheet. The JSON carries the licence, the source and a description of every column inside the file itself, so a copy that has been passed on twice still says where it came from.
The word list
All 11,000 words of the HSK 3.0 vocabulary, level by level. Levels 7, 8 and 9 are examined as one band and the syllabus publishes them as one list of 5,600; the division into three of about 1,867 is Gushi’s own, so that there is a next level to reach.
| Level | Rows | CSV | JSON |
|---|---|---|---|
| All levels | 11,000 | CSV 1.7 MB | JSON 3.9 MB |
| HSK 1 | 300 | CSV 46 kB | JSON 108 kB |
| HSK 2 | 200 | CSV 34 kB | JSON 76 kB |
| HSK 3 | 500 | CSV 89 kB | JSON 191 kB |
| HSK 4 | 1,000 | CSV 173 kB | JSON 375 kB |
| HSK 5 | 1,600 | CSV 274 kB | JSON 596 kB |
| HSK 6 | 1,800 | CSV 275 kB | JSON 637 kB |
| HSK 7 | 1,867 | CSV 258 kB | JSON 633 kB |
| HSK 8 | 1,867 | CSV 251 kB | JSON 626 kB |
| HSK 9 | 1,866 | CSV 250 kB | JSON 625 kB |
The character list
The 3,088 characters HSK 3.0 asks you to read (认读字). The smaller set of 1,200 you have to write by hand is a subset of these.
| Level | Rows | CSV | JSON |
|---|---|---|---|
| All levels | 3,088 | CSV 461 kB | JSON 1.1 MB |
| HSK 1 | 246 | CSV 35 kB | JSON 86 kB |
| HSK 2 | 125 | CSV 20 kB | JSON 47 kB |
| HSK 3 | 284 | CSV 46 kB | JSON 104 kB |
| HSK 4 | 441 | CSV 73 kB | JSON 163 kB |
| HSK 5 | 431 | CSV 76 kB | JSON 164 kB |
| HSK 6 | 413 | CSV 67 kB | JSON 151 kB |
| HSK 7 | 383 | CSV 49 kB | JSON 127 kB |
| HSK 8 | 383 | CSV 47 kB | JSON 125 kB |
| HSK 9 | 382 | CSV 48 kB | JSON 126 kB |
The grammar syllabus
The 428 grammar points published here, each with the pattern the syllabus prints beside it, one sentence on what it does, and an example. Levels 7 to 9 are one file, exactly as the syllabus publishes them.
| Level | Rows | CSV | JSON |
|---|---|---|---|
| All levels | 428 | CSV 96 kB | JSON 182 kB |
| HSK 1 | 45 | CSV 9 kB | JSON 20 kB |
| HSK 2 | 48 | CSV 10 kB | JSON 22 kB |
| HSK 3 | 59 | CSV 12 kB | JSON 26 kB |
| HSK 4 | 67 | CSV 14 kB | JSON 29 kB |
| HSK 5 | 64 | CSV 15 kB | JSON 30 kB |
| HSK 6 | 49 | CSV 12 kB | JSON 23 kB |
| HSK 7–9 | 96 | CSV 24 kB | JSON 45 kB |
Hear about it when the data changes
The syllabus does not move often, but the data does: a corrected meaning, a missing example, a new recording. Leave an address and we write once, when there is something to say.
What is in the columns
The same columns in both formats, in this order. The JSON repeats these descriptions inside every file, so a file that has been passed on twice still explains itself.
level- The HSK 3.0 level, 1 to 9.
hanzi- The word or character in simplified Chinese.
pinyin- Pinyin with tone marks.
tones- One tone number per syllable, space separated. 5 is the neutral tone.
meaning- The English meaning. Several senses are separated by a semicolon.
exampleChinese, examplePinyin, exampleEnglish- One example sentence, in all three.
audio- The address of a recording of the word, or empty where there is none.
point, pattern, explanation- Grammar files only: the point as the syllabus names it, the pattern printed beside it, and one sentence on what it does.
categoryChinese, categoryEnglish- Grammar files only: the section of the syllabus the point belongs to.
An audio address looks like /audio/az-5f97.mp3, where the hexadecimal is the character itself. The recordings are free to play from this site and are the one column the licence does not cover: see below. Please do not hotlink them from a product.
Use it, credit it
The data is published under Creative Commons Attribution-ShareAlike 4.0. You may copy it, change it, build something on it and sell what you build, on two conditions: name Gushi and link back, and pass on anything you derive from it under the same licence.
Share-alike is not a preference on our part. The English meanings are adapted from CC-CEDICT, which is CC BY-SA itself, so this data could never have gone out under anything more permissive.
The recordings the audio column points at are the one thing outside all of this. They are synthesised speech; anyone may play them from this site, but they are not ours to hand on as a corpus, so the licence does not cover them.
The credit we ask for, ready to paste:
<a href="https://gushichinese.com/hsk-3.0/downloads">HSK 3.0 syllabus data by Gushi</a>, CC BY-SA 4.0Where it comes from: the lists are the official 考试大纲 for HSK 3.0, published November 2025. The English meanings are adapted from CC-CEDICT (CC BY-SA 4.0), the example sentences come largely from the Tatoeba Project (CC BY 2.0 FR), and the selection, pinyin, tone numbers and checking are ours. Data last checked 2026-10-07.