Sources & accuracy
HSK Prep's word lists are only useful if they are right. This page lists where every word comes from and the checks it passes before each release.
The official syllabus
HSK 3.0 is based on the Chinese Proficiency Grading Standards for International Chinese Language Education (国际中文教育中文水平等级标准, GF 0025-2021), issued by the Ministry of Education of the PRC and the State Language Commission and in force since 1 July 2021 (MOE announcement). Its vocabulary appendix lists 11,092 words and its character appendix 3,000 characters, divided into levels 1–6 and a combined 7–9 band.
Data used to build and check the lists
- Primary list: ivankra/hsk30 (MIT license), a transcription of the GF 0025-2021 word, character and grammar appendices, reconciled with the official HSK word query data.
- Check 1, the PDF: elkmovie/hsk30, an independent OCR transcription of the official PDF, used to confirm every headword and its position.
- Check 2, the official query system: HSK official query system data (words and pinyin as published on chinesetest.cn), used to confirm every reading.
- Check 3, a dictionary: CC-CEDICT (CC BY-SA 4.0, release of 2026-10-19), used to confirm each word's reading and each character's reading, and as the source of English meanings for most words from HSK 3 up.
Automated checks (all passing)
| C1 word counts per level = GF 0025-2021 | 1: 500, 2: 772, 3: 973, 4: 1000, 5: 1071, 6: 1140, 7-9: 5636 |
| C2 headword + position match official PDF (OCR) | 11092 compared |
| C3 pinyin = official chinesetest.cn query-system data | 11092 compared, 2 documented source typos |
| C4 syllables valid + tone mark on the correct vowel | 22081 syllables |
| C5 syllable count aligns with characters (erhua aware) | all forms |
| C6 whole-word reading agrees with CC-CEDICT | 138 neutral-tone convention differences (official list kept), 6 documented colloquial readings |
| C7 each syllable is a dictionary reading of its character | all syllables |
| C8 3,000 characters, 300/300/300/300/300/300/1200, match PDF table | 3000 characters |
Last checked 2026-10-19. The build stops if any check fails or any new difference appears that isn't documented below.
Pinyin conventions
- Tone changes of 一 and 不 are written as spoken (一起 yìqǐ, 不错 búcuò), as on the official query site. The dictionary form (yīqǐ, bùcuò) is kept and shown in the app. 169 words are affected.
- Third-tone changes (你好 nǐhǎo said as níhǎo) are not written, following standard pinyin.
- Light (neutral) tones follow the official list. Where CC-CEDICT differs only in a light tone (138 words, e.g. 回来 huílai/huílái), the official reading is kept.
- Separable words such as 见面 jiàn∥miàn are marked in the app.
English meanings
English meanings for all of HSK 1 and HSK 2, and for 540 harder words where a dictionary's first sense is misleading or missing, were written by hand for the sense the syllabus intends. The remaining 9,280 meanings are selected automatically from CC-CEDICT by part of speech, skipping variant, literary and rare senses. If you spot a meaning that could be better, please tell us.
Documented differences between sources
Each case below was reviewed individually against the official standard.
| Word | Resolution |
|---|---|
| L1-0224 | Query-system data lists 妈妈 without the 妈 variant; the official PDF prints 妈妈｜妈. PDF wins. |
| L4-0338 | Query-system data has 会 (huìlǜ) in place of 汇率 — a typo in that dataset. The PDF prints 汇率 huìlǜ. |
| 会 | Query-system data has an extra 会 because of the 汇率 typo above. |
| L1-0325 | 谁 shuí is the official alternative reading (shéi/shuí); CC-CEDICT only lists shei2 for this headword. |
| L2-0500 | 熟 shóu is the official colloquial reading (shú/shóu). |
| L3-0798 | 血 xiě is the colloquial reading listed by the standard; CC-CEDICT gives xue4 for the bare character. |
| L7-0552 | 出血 chūxiě follows 现代汉语词典 (colloquial 血 xiě). |
| L7-4508 | 献血 xiànxiě follows 现代汉语词典. |
| L7-4847 | 钥匙 yàoshi is correct; the per-character table in CC-CEDICT only knows 钥 as yue4 in another entry. |
| L7-1836 | The OCR file numbers 火药 as 1856 (misread digit); it sits between 1835 火焰 and 1837 伙食, so its position is 1836 as in the PDF. |
Speaking prompts
The speaking practice sets are original material written for HSK Prep in the format of the official HSK 3.0 sample papers. They are not real exam questions. Every Chinese sentence with pinyin is checked automatically: syllables align with characters, and every syllabus word inside a prompt carries its official reading.
Licenses
Word-list data: GF 0025-2021 (public national standard); ivankra/hsk30 (MIT). English meanings partly from CC-CEDICT, © MDBG and contributors, licensed under CC BY-SA 4.0; the combined glossary file /data/hsk*.json is shared under the same license.