How long does it take to learn Korean?

The milestones on the way to Korean fluency

Korean is FSI Category IV, and the 2,200-hour figure is the estimate to professional working proficiency, not the point where the language becomes usable. The waypoints come much earlier:

FSI publishes a total to General Professional Proficiency and no figure for any earlier level, so the A2 and B1 waypoints here are our own rounded shares of that total. They are shown so this page can answer when conversation becomes possible without inventing an FSI figure that does not exist.

FSI measures its endpoint as Speaking 3 and Reading 3 on the Interagency Language Roundtable scale, which published crosswalks place somewhere around CEFR B2 to C1. Those crosswalks disagree with each other and no official conversion exists, so treat the CEFR labels here as an approximate guide rather than an equivalence.

Studying more or less than an hour a day? Model your own pace with the fluency calculator.

How many hours does it take to become fluent in Korean?

The US Foreign Service Institute puts Korean at 2,200 class hours to professional working proficiency, the level at which you can work in the language. It publishes that as 88 weeks of full-time training, on a heavier schedule than its shorter courses: the published totals average 25 class hours a week. Fluency has no single official definition, which is why the hour count is the honest way to answer this.

How long does it take the average person to learn Korean?

There is no measured figure for the average person, and anyone quoting one is guessing. The 2,200-hour benchmark describes full-time students in small classes with several hours of individual study a day, which is not how most people learn. The hour total is still the useful anchor: at an hour a day it works out to about 73 months, and at half an hour a day about 147 months.

How long does it take to learn Korean on your own?

The Foreign Service Institute does not publish a self-study equivalent, so the "double it if you are on your own" numbers you see are unsourced. What is on the record is that the 2,200 hours were measured under intensive instruction with a teacher. Treat them as a benchmark set under ideal conditions rather than a forecast for solo study, and track your progress in hours logged instead of months elapsed.

How long does it take to learn Korean from English?

The 2,200-hour estimate is specifically for native English speakers. The FSI sorts languages by how far they sit from English and places Korean in Category IV, the hardest group it publishes. If you already speak a language closely related to Korean, you would expect to move faster than the benchmark, though the FSI publishes no figure for that case.

How fast can you learn Korean?

The realistic answer is a pace rather than a date. 2,200 hours is about 147 months at half an hour a day, about 73 months at an hour a day, and about 105 weeks at three hours a day. That is the same amount of learning on three different calendars, so the variable you actually control is how many hours a week you can keep logging.

Can you learn Korean faster by watching K-dramas?

Watching does not shrink the 2,200 hours, but it changes whether you keep logging them. Graded, comprehensible viewing builds listening and vocabulary in context, and saving the words you meet for spaced review turns watch time into study time. There is a catch worth knowing before betting the plan on television. Across the 106 title and language entries in our open Netflix difficulty index, none is rated A1 and only 2 sit at A2, while 82 percent sit at B1 or B2. Ordinary adult television starts at intermediate, so the early hours still need graded or beginner material and the catalogue opens up later. Two honest limits on that: it describes our catalogue rather than Netflix, since we chose these titles as plausibly useful for learners and the selection shapes the distribution as much as the material does, and the bands are editorial estimates of listening difficulty rather than learner test results. That is the approach LingoBinge takes on Netflix, swapping a few subtitle words to your level while you watch. It grades Indonesian today, and Korean is on the waitlist.

How long until you can hold a conversation in Korean?

Basic conversation (CEFR A2) arrives far before the full total: around 660 study hours for Korean, which is roughly 22 months at an hour a day. FSI hours are intensive classroom hours for motivated adult learners, so they are a benchmark, not a promise.

What learning Korean actually involves

The hardest part of Korean for English speakers

The hardest system is honorification, specifically the split between subject honorification, keyed to the grammatical subject, and addressee honorification, keyed to the hearer. Kwon and Shin describe subject honorification as socio-pragmatic in interpretation but syntactically constrained, and attribute learner error to its low cue validity and its frequent co-occurrence with pragmatically salient features, so the suffix gets read as a general politeness marker rather than as agreement. In self-paced reading, English-speaking learners detected semantic anomalies but not morphosyntactic violations.

Source: Kwon, Nayoung and Gyu-Ho Shin (2025), 'Overgeneralization of Korean subject honorification by English-speaking learners of Korean', Bilingualism: Language and Cognition, FirstView, pp. 1-17, published online 9 December 2025

What is easier in Korean than people expect

The Korean predicate inflects for honorification but never for person. WALS records no verbal person marking, so one form serves a first, second or third person subject, singular or plural, and it records no definite or indefinite article, so nothing corresponding to English the or a has to be selected in front of a noun. The conjugation tables that dominate Romance and Slavic study have no counterpart here.

Sources: Anna Siewierska, Verbal Person Marking (feature 102A), Korean datapoint, value 'No person marking', citing Sohn 1994, in Matthew S. Dryer and Martin Haspelmath (eds.), The World Atlas of Language Structures Online, Max Planck Institute for Evolutionary Anthropology, Leipzig, 2013; Matthew S. Dryer, Definite Articles (feature 37A), Korean datapoint, value 'No definite or indefinite article', citing Sohn 1999, in Matthew S. Dryer and Martin Haspelmath (eds.), The World Atlas of Language Structures Online, Max Planck Institute for Evolutionary Anthropology, Leipzig, 2013; Kwon, Nayoung and Gyu-Ho Shin (2025), 'Overgeneralization of Korean subject honorification by English-speaking learners of Korean', Bilingualism: Language and Cognition, FirstView, pp. 1-17, published online 9 December 2025

Why the FSI puts Korean where it does

The Foreign Service Institute assigns Korean 88 weeks and 2,200 class hours in Category IV, its super-hard band, against roughly 44 weeks and 1,012 hours for Polish in Category III, so Korean takes twice the weeks and more than twice the classroom time. Korean reaches that band with no verbal person marking, no articles and an alphabetic script of 19 consonant and 21 vowel letters. FSI publishes no per-language rationale, so its bands reflect time observed at FSI to reach an ILR speaking and listening score of 3 rather than a grammar checklist.

Sources: National Institute of Korean Language, Hangeul Matchumbeop (Korean Orthography), Notice No. 2017-12, in force 28 March 2017, Chapter 2 Article 4 with Appendices 1 and 2, in the Korean Language Norms database; Anna Siewierska, Verbal Person Marking (feature 102A), Korean datapoint, value 'No person marking', citing Sohn 1994, in Matthew S. Dryer and Martin Haspelmath (eds.), The World Atlas of Language Structures Online, Max Planck Institute for Evolutionary Anthropology, Leipzig, 2013; Matthew S. Dryer, Definite Articles (feature 37A), Korean datapoint, value 'No definite or indefinite article', citing Sohn 1999, in Matthew S. Dryer and Martin Haspelmath (eds.), The World Atlas of Language Structures Online, Max Planck Institute for Evolutionary Anthropology, Leipzig, 2013; U.S. Department of State, Foreign Service Institute, National Foreign Affairs Training Center, 'Foreign Language Training', language learning timelines by category

Checked and not published

Claims about Korean that circulate widely but that we could not trace to a primary source, so this page does not make them.

• Any Sino-Korean-origin share of the lexicon. A widely repeated proportion was deliberately not stated: no primary published count was located, and any such figure varies with whether the base is a dictionary headword list or running text.
• A count of Korean addressee honorific speech levels. Search snippets quoting the Cambridge Handbook and an MDPI article assert six (hapsyo, hao, hakey, hayla, hayyo, hay), but no primary text stating the count could be opened, so no number is claimed anywhere in the six fields.
• The traditional Korean school-grammar case count (commonly seven case-particle classes) as a contrast to the WALS '6-7 cases' bucket. The brief asks for both counts where they differ, but the traditional figure could not be traced to a primary source, so only the WALS bucket is stated and is attributed to WALS by name.
• 'Korean has no grammatical gender.' Deliberately NOT written. WALS 44A (Gender Distinctions in Independent Personal Pronouns) records Korean as '3rd person singular only', which a flat no-gender line would have contradicted. That feature measures pronoun gender rather than noun class, so the safe move was to drop the claim entirely; pleasantSurprise rests on honorification, articles and person marking instead.
• The segmental form of the subject honorific suffix (-si-). Kwon and Shin's abstract names only 'the subject honorific suffix' without giving the form, so hardestThing uses the functional term and never cites the morpheme shape.
• Any time estimate for learning Hangul (an hour, an afternoon, a day). No source publishes one, so no such claim appears in writingSystem or fsiNote.
• The claim that the addressee honorific is carried by an obligatory sentence-final ending on every finite predicate. Widely repeated and probably true, but the only accessible statement of it was a search snippet, so grammarLoad says only that honorification runs on two tracks, one keyed to the subject and one to the addressee.
• A total letter count for Hangul. The primary orthography does state one in Article 4, and it verifies, but the figure is on the brief's forbidden list because it is the misleading half of the fact: the same article's Appendix 2 sets the working dictionary inventory at 19 consonant and 21 vowel letters. Only the 19 and 21 counts, and the 5 doubled and 11 compound letters derived in Appendix 1, are stated.
• The characterisation of Korean case markers as detachable particles rather than affixes. WALS 51A-kor labels them 'Case suffixes', so the particle framing was dropped; grammarLoad states the suffix analysis and the separate 20A-kor 'Exclusively concatenative' value instead.
• Any claim about Polish grammar. Polish is used solely as an FSI hours comparator, and nothing beyond its Category III placement on the state.gov page was sourced.