Skip to main content

JLPT Vocabulary by Level: How Many Words for N5 to N1

JLPT vocabulary by level is roughly N5 ~800 words, N4 ~1,500, N3 ~3,750, N2 ~6,000, and N1 ~10,000 on a cumulative basis. In other words, each level assumes everything below it.12345 Every figure is an unofficial estimate, because the test stopped publishing an official vocabulary list in 2010.6 This hub gives you the headline table, explains the common mistake of estimating "new words per level" by subtraction, and points you to the right level list and study method.

Why There Is No Official JLPT Vocabulary List

There is no official, published list of JLPT vocabulary for any level.6 That is why every count you see online is hedged, and why reconstructed lists disagree with one another.

The old 出題基準 lists (1994–2010)

Before the 2010 revision, the test content was governed by the 出題基準しゅつだいきじゅん (Shutsudai Kijun, "Test Content Specification"), first published in 1994 and revised in 2004.7 It contained kanji, expression, vocabulary, and grammar lists for all four levels of the old test.7

The specification was a reference for test writers, not a study syllabus for candidates. Roughly 20% of any exam's kanji, vocabulary, and grammar could be drawn from outside the listed items at the compilers' discretion.7

The old test had four levels: 4-kyū was the lowest, and 1-kyū was the highest.7 The approximate vocabulary targets stated for that era were 4-kyū ~800 words (~100 kanji), 3-kyū ~1,500 words (~300 kanji), 2-kyū ~6,000 words (~1,000 kanji), and 1-kyū ~10,000 words (~2,000 kanji).7

The headline numbers are inherited from the old four-level test

The familiar ~1,500 / ~6,000 / ~10,000 figures are not new estimates. They are the old 3-kyū, 2-kyū, and 1-kyū vocabulary targets from the 出題基準 era, carried forward to the new levels.7

The 2010 five-level test maps roughly onto the old four levels this way: N5 is the same as old 4-kyū, N4 is the same as old 3-kyū, N3 sits between old 3-kyū and old 2-kyū, N2 is the same as old 2-kyū, and N1 is slightly more advanced than old 1-kyū but at the same passing level.7 N3 is the genuinely new tier. It was inserted to bridge the large gap between old 3-kyū and old 2-kyū.7

What changed in the 2010 revision

In 2010, the test was restructured from four levels to five (N1 through N5).87 At the same time, the organizers stopped publishing the Test Content Specification, and it has not been published since.6

The stated reason is a shift in testing philosophy toward communicative competence: the ability to use the language in real communication. The organizers hold that the goal of studying Japanese is to communicate, not to memorize vocabulary, kanji, and grammar items. They therefore judged that publishing a specification listing those items was not necessarily appropriate.6

In place of the discontinued lists, the organizers publish a "Summary of Linguistic Competence Required for Each Level," the composition of test items, and sample questions for every level.6 Wikipedia summarizes the same change this way: no Test Content Specification is published, because studying from kanji and vocabulary lists is discouraged.7

So where do the numbers come from

Because no official post-2010 list exists, every per-level vocabulary count is a community reconstruction. These reconstructions are built from remembered exam items, the old 出題基準, and difficulty estimates.7910 Tanos, jpdb, and JLPT Sensei are three of the better-known efforts.910

These reconstructions differ because they use different source corpora and counting methods. That spread is why every figure in this article carries a "~" hedge.910

How Many Words Per Level: The Cumulative Table

The widely cited cumulative figures

The table below gives the commonly cited cumulative vocabulary targets. Each level assumes everything below it, and every figure is an unofficial estimate. The right-hand column points to the full word list for each level.

LevelCumulative vocabulary (unofficial estimate)Full word list
N5~8001JLPT N5 Vocabulary List
N4~1,5002JLPT N4 Vocabulary List
N3~3,7503JLPT N3 Vocabulary List
N2~6,0004JLPT N2 Vocabulary List
N1~10,0005JLPT N1 Vocabulary List

The N4, N2, and N1 cumulative figures (~1,500 / ~6,000 / ~10,000) match the old-level targets from the 出題基準 era, which is their historical origin.7

Read these as planning tools, not pass marks

The JLPT is not scored against a vocabulary checklist. It scores language knowledge (vocabulary and grammar), reading, and listening. There is no fixed vocabulary cutoff to "complete."11

The official position is that the goal is communicative use of the language, not memorizing a fixed list. That is why no list is published as a syllabus.6 A cumulative count is useful for scoping study time, not for telling you whether you will pass.

General word-count theory asks how many words you need for broad comprehension. That is separate from these test-specific figures and is covered in the article on how many words you need.

Cumulative vs. New Per Level: The Number That Trips Everyone Up

Why subtraction overstates the new layer

You cannot reliably find the genuinely new vocabulary at a level by subtracting one cumulative headline from another. The cumulative totals (~800 / ~1,500 / ~3,750 / ~6,000 / ~10,000) are rounded estimates, and across sources they come from different reconstructions.7910

They are not built on one consistent base. When you subtract two independently reconstructed totals, you double-count overlap and inflate the new layer.7910

Naive subtraction gives ~2,250 new words at N2 (6,000 − 3,750), or a commonly quoted ~3,000 (6,000 − 3,000), and ~4,000 new words at N1 (10,000 − 6,000).10 Direct counts of the per-level new-words lists give substantially smaller figures.910

What the level lists actually add

A more reliable basis for "new per level" is a direct count of distinct new entries. Tanos publishes per-level lists that contain only that level's new words. jpdb publishes cumulative counts, and the difference between successive levels gives the new layer.910 The two methods are shown side by side below.

LevelNew words this level (direct count, band)Tanos new-words list9jpdb cumulative, then delta10
N5~670–800 (the base layer)~689667 (base)
N4~575–700 new~6021,242, so +575
N3~1,700–1,800 new~1,8002,976, so +1,734
N2~1,750 new (not ~3,700)per-level new list4,792, so +1,816
N1~2,700–3,300 new (not ~4,000)per-level new list8,124, so +3,332

The Tanos per-level lists are new-words-only. Each level's list does not repeat lower-level vocabulary. The N5 list is ~689 words and the N4 list is ~602 words.9

The jpdb cumulative counts are N5 667, N4 1,242, N3 2,976, N2 4,792, and N1 8,124 unique words.10 Their successive differences are N4 +575, N3 +1,734, N2 +1,816, and N1 +3,332.10

The figures are given as bands because Tanos and jpdb differ by hundreds at the upper levels. Both methods agree on the conclusion that matters: the new layer at each level is far smaller than cumulative subtraction implies.910

N2 adds far fewer new words than the headline gap suggests

If you read the cumulative jump 6,000 − 3,750 as "~2,250 new" (or the looser 6,000 − 3,000 as "~3,000 new"), you overstate N2. A direct count puts the genuinely new N2 layer near ~1,750, not ~3,700.10

What this means for your study plan

The jump from N3 to N2 adds a smaller new layer (~1,750 by direct count, jpdb delta +1,816) than the cumulative headline of ~2,250 suggests.10 To cover that layer in practice, see "JLPT N2 Kanji and Vocabulary Strategy".

N3 itself is one of the heaviest single jumps, at ~1,700–1,800 new words. That fits N3's role as the bridge tier inserted between old 3-kyū and old 2-kyū.710

For the breakdown by category and register within a level, open that level's own list. That is where the per-category counts live.

How to Use These Counts

Pick your level, then open its list

A count tells you the rough size of the job. The breakdown by category and register lives in each level's list. Pick the level you are targeting, then open the matching article: JLPT N5 Vocabulary List, JLPT N4 Vocabulary List, JLPT N3 Vocabulary List, JLPT N2 Vocabulary List, or JLPT N1 Vocabulary List.

Don't drill the list, acquire the words

The official guidance frames the test as measuring communicative competence rather than rote memorization of a list.6 That favors learning words in context over drilling a list front to back.

A count tells you scope, not method. High-frequency-first study and reading-based learning are covered in the articles on word-frequency coverage, sentence mining, and your vocabulary acquisition strategy.

Pair a level-mapped SRS deck with a frequency-first method

To work one level at a time, J-Compass recommends Amenokori. It maps its vocabulary, grammar, and kanji decks to N5 through N1 and schedules them with FSRS, so advancing a level is as simple as opening the next deck.1213 Pair that structure with a frequency-first reading habit. See how it stacks against other tools in the guide on choosing your resources.1213

Good to know

Counts vary by source, and that's expected

Because the lists are reconstructions, sources disagree by hundreds of words at the same level. For N3 cumulative vocabulary, JLPT Sensei gives ~3,7503 while jpdb counts 2,976.10 For N1, JLPT Sensei gives ~10,0005 while jpdb counts 8,124.10

Treat any single number as an estimate, not a fixed fact.

Vocabulary is only one of three skills tested

The JLPT measures language knowledge (vocabulary and grammar), reading, and listening together.11 A large vocabulary helps all three, but passing is not a vocabulary checklist. The exam's mechanics, scoring, and section structure are covered separately in the JLPT overview. This hub stays on counts.

"Words" is a fuzzy unit

Counts differ on whether conjugations, compounds, and katakana variants are treated as separate words. Lemma-versus-surface-form choices alone can shift a total by hundreds. In other words, counts change depending on whether you count a dictionary headword or each form as it appears in text. The gap between the Tanos new-words-only lists9 and the jpdb cumulative counts10 at the same level shows this. It is another reason every figure here is approximate.

See also

References

Footnotes

  1. JLPT Sensei. "JLPT N5 Vocabulary List." https://jlptsensei.com/jlpt-n5-vocabulary-list/ 2

  2. JLPT Sensei. "JLPT N4 Vocabulary List." https://jlptsensei.com/jlpt-n4-vocabulary-list/ 2

  3. JLPT Sensei. "JLPT N3 Vocabulary List." https://jlptsensei.com/jlpt-n3-vocabulary-list/ 2 3

  4. JLPT Sensei. "JLPT N2 Vocabulary List." https://jlptsensei.com/jlpt-n2-vocabulary-list/ 2

  5. JLPT Sensei. "JLPT N1 Vocabulary List." https://jlptsensei.com/jlpt-n1-vocabulary-list/ 2 3

  6. Japan Educational Exchanges and Services (JEES) and the Japan Foundation. "FAQ: Studying for the test." Official JLPT website. https://www.jlpt.jp/e/faq/index.html 2 3 4 5 6 7

  7. "Japanese-Language Proficiency Test." Wikipedia (English), "Test Content Specification," "Levels," and "Change in 2010" sections. https://en.wikipedia.org/wiki/Japanese-Language_Proficiency_Test 2 3 4 5 6 7 8 9 10 11 12 13 14 15

  8. Japan Educational Exchanges and Services (JEES) and the Japan Foundation. "Topics 2010" and overall pass-mark notice. Official JLPT website. https://www.jlpt.jp/e/topics/list2010.html

  9. Tanos (Jonathan Waller). "JLPT Resources: Free Japanese Vocabulary lists and MP3 sound files." Per-level vocabulary lists (new-words-only, not cumulative). https://www.tanos.co.uk/jlpt/ 2 3 4 5 6 7 8 9 10 11

  10. jpdb.io. "Japanese Language Proficiency Test (Vocabulary list)," cumulative per-level word counts. https://jpdb.io/vocabulary-list/924/japanese-language-proficiency-test 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

  11. Japan Educational Exchanges and Services (JEES) and the Japan Foundation. "Summary of Linguistic Competence Required for Each Level." Official JLPT website. https://www.jlpt.jp/e/about/levelsummary.html 2

  12. Amenokori. Product landing page. https://amenokori.com 2

  13. Amenokori. Mobile-app page. https://amenokori.com/mobile-app/ 2