How long does it take to learn Mandarin?

The best-known answer is 88 weeks of full-time study, the figure the US Foreign Service Institute gives for Mandarin to reach professional working ability in speaking and listening — the longest of any language it rates. But for Mandarin a single number answers the wrong question, because there is no single finish line. Understanding speech, reading and writing by hand each run on their own clock, and China's own proficiency standard shows how far apart they are: its top band asks for 11,092 words, 3,000 characters you can recognise, and only 1,200 you can write.

Updated 7 October 2026 · 8 min read

The 88-week figure, and what it measures

The Foreign Service Institute (FSI) trains American diplomats and publishes how long its students usually take to reach an integrated score of 3 (Speaking + Listening) on the ILR scale. It sorts languages into four categories, and Mandarin is in the last of them, which the page itself calls "super-hard languages":

Category IV is a short list: Arabic, Cantonese, Mandarin, Japanese and Korean. A typical FSI week is 23 hours in class and 17 of self-study, so 88 weeks of both comes to more than three thousand hours of work — for adults chosen partly for aptitude, studying full-time with trained teachers.

Notice what the target leaves out. The score is for speaking and listening. It says nothing about reading a newspaper or writing a note by hand, and for most languages that hardly matters, because once you can say a word you can very nearly read it. In Mandarin it matters a great deal.

Three clocks, not one

A learner of Mandarin is really keeping three schedules at once:

• Listening and speaking run on words and tones. Characters play no part: you can hold a conversation without knowing a single one.
• Reading needs the same words, plus the characters they are written in. A word you can say is not a word you can read.
• Handwriting needs you to produce those characters from memory, stroke by stroke, which is a much smaller and slower skill than recognising them.

Chinese even has everyday phrases for the gap. 听不懂 tīng bu dǒng is "I can't understand it" when you hear something; 看不懂 kàn bu dǒng is the same thing for something you read. A learner meets both, usually at different times, and progress on one tells you little about the other.

China's own milestones: the HSK standard

The clearest picture of the three clocks comes from China's Ministry of Education. Its 2021 standard for teaching Chinese to foreign learners — Chinese Proficiency Grading Standards for International Chinese Language Education, GF 0025-2021, the basis of the "HSK 3.0" exams — divides learning into three bands and nine levels, and puts a number on everything:

The figures are cumulative. Two things stand out.

The handwriting clock is set deliberately slow. At every band the standard expects you to write only a fraction of what you can read: 300 of 900, 700 of 1,800, 1,200 of 3,000. That is not a concession to weak students: the standard treats recognising a character and producing one from memory as separate skills, with separate lists.

Words keep coming long after characters slow down. Level by level, the standard adds characters at an even pace and words at an accelerating one:

By the end, every character is doing the work of nearly four words. That is the long-run shape of learning Mandarin: the characters are a large but finite job front-loaded into the first years, and vocabulary is the job that never quite ends.

What those milestones buy you in real Chinese

Counts on a syllabus are a promise; coverage is what you actually get. Measured on 85 million words of Chinese film and television dialogue (OpenSubtitles 2018, close to how people talk), here is what the nearest measured point to each HSK band covers — counting characters for the reading clock and words for the listening one:

The two clocks finish at completely different times. By the end of the elementary band almost every character on a screen of subtitles is one you have met, yet a large share of the words in the same dialogue are still new, because known characters combine into unknown words. By the advanced band the character clock has effectively stopped, while the word clock is still short of 90%. Why the two curves differ so much is the subject of how many words you need to speak Mandarin.

This is also why the FSI number is so large. Measured on equal-sized samples, Mandarin's 10,000 most common words cover 88.6% of its dialogue; Spanish reaches the same point with 5,607 words. A listener of Mandarin needs a far bigger vocabulary for the same comprehension, and the hardest languages to learn sets that beside every other language.

What it looks like a day at a time

The FSI's weeks assume full-time study. For a learner doing one lesson a day, the more useful measure is lessons. Verbavia (our own course) introduces 6 new Mandarin words a lesson, in order from a 6,000-word frequency list, and moves through the CEFR levels on a fixed schedule. Done one a day, the levels arrive like this:

The first two columns are for the full-length lesson, about 20–30 minutes; the shorter lengths spread each step of the syllabus over more days rather than skipping any of it. A level also only moves on after a lesson that went well, so these are the fastest the schedule allows. The first 704 lessons, up to the start of C2, are written by hand, and the very first is on pinyin, so the listening clock starts before any characters are asked of you.

Set beside the HSK table, the arithmetic is sobering in a useful way: the elementary band's 2,245 words take more than a year at six a day, and the advanced band's 11,092 are beyond what any daily course alone gets through. The words after the first few thousand come mostly from reading and listening to real Chinese, which is why graded reading and Chinese-subtitled video matter so much later on.

So how long does Mandarin take?

It depends on which clock you are watching.

• To hold a basic conversation: the fastest clock. The grammar is light, 161 words cover half of all dialogue, and A2 begins after 113 daily lessons; the real work is in the four tones, which need practice from the first day.
• To follow ordinary speech comfortably: this is what the FSI measures, and its 88 weeks of full-time study, more than three thousand hours with self-study, is the honest scale for an adult in a classroom.
• To read: the characters are a finite job — the standard puts 1,800 at the end of the intermediate band, and 2,000 already cover 97.3% of the characters in dialogue — but the words they spell out take as long as listening does.
• To write by hand: the slowest clock, and the one the official standard itself sets lowest. Most learners can let it lag.

Whether that makes Mandarin hard, and why speaking is easier than its reputation, is in is Mandarin hard to learn. If you are choosing between Chinese languages, Mandarin vs Cantonese compares the two in the FSI's top category.

The most common Mandarin words are free, in frequency order and with pinyin, and the two-minute Mandarin vocabulary test estimates how far along the word clock you already are. The Mandarin course is free on the web and works in any browser, including on a phone.

Sources

• Foreign Service Institute, National Foreign Affairs Training Center: language learning timelines, checked 2026-09-25; read for this post from the Internet Archive copy of 2 October 2026 — Category IV, "super-hard languages", 88 weeks (2,200 class hours), and the 23 + 17 hour week.
• Ministry of Education of the People's Republic of China and the State Language Commission: announcement of the Chinese Proficiency Grading Standards for International Chinese Language Education (GF 0025-2021, in force from 1 July 2021), and the standard itself — the summary table of quantitative indicators, and sections 3.1–3.3 for the handwritten-character counts.
• Wikipedia: Hanyu Shuiping Kaoshi — the November 2025 HSK 3.0 exam syllabus and its rollout.
• Wikipedia: List of Frequently Used Characters in Modern Chinese — the 3,500-character list.
• Lison, P. & Tiedemann, J. (2016). OpenSubtitles2016: Extracting Large Parallel Corpora from Movie and TV Subtitles. Proceedings of LREC 2016. aclanthology.org
• Dave, H. FrequencyWords: frequency lists from OpenSubtitles 2018 (zh_cn), content licensed CC BY-SA 4.0. github.com. Every coverage figure is computed by scripts/frequency_curves.py and printed from its output rather than typed.
• The course figures come from Verbavia's own course files: 6 new words a lesson, the level boundaries, the lesson lengths and the hand-written lesson count.