JLPT N1 Vocabulary List: ~10,000 Words and Why You Read Instead of Drill
The JLPT N1 vocabulary list is the cumulative ~10,000-word recognition vocabulary expected at N1, the top level of the Japanese-Language Proficiency Test.1 It is also the level where the "list" is least useful. The words genuinely new at N1 form a long tail of low-frequency literary, formal, and specialized vocabulary. You meet them by reading, not by drilling.2
Overview
There is no official JLPT vocabulary list. The Japan Foundation and JEES do not publish a "Test Content Specification" that lists vocabulary, kanji, or grammar items. They set this policy when the current five-level test (N1–N5) was adopted, beginning with the July 2010 test.3 Every "N1 list" in circulation is therefore a community reconstruction of the post-2010 scope, not a document issued by the exam body.
The widely cited ~10,000 figure is the cumulative recognition vocabulary for N1. It is community consensus, not an official count. Tanos summarizes the N1 requirement as roughly 10,000 words;4 JLPTsensei states a learner needs "about 10,000 words" to pass.1 Both rest on durable reconstructions, not on a list attributed to JEES. JLPT Vocabulary by Level: How Many Words for N5 to N1 frames the unofficial cumulative counts across all five levels and the cumulative-vs-new-layer model this page sits at the top of.
N1 vocabulary is cumulative. The ~10,000 total already includes the N2 base, which itself contains the N3, N4, and N5 layers, so this article does not re-list those words. It foregrounds the layer new beyond N2 and points you to the JLPT N2 Vocabulary List for the prerequisite base, with the JLPT N3 Vocabulary List, JLPT N4 Vocabulary List, and JLPT N5 Vocabulary List as the foundation beneath it.
The official N1 can-do bar is reading "newspaper editorials and critiques" and "writings with logical complexity and/or abstract writings on a variety of topics," and following lectures at natural speed.2 That bar describes broad written and abstract register, not a bounded everyday vocabulary. This is why a flat N1 list is the least drillable list of any level.
The ~10,000 figure is the cumulative total. This article catalogs only the words new beyond N2 and organizes them by register. That is why the N2/N3/N4/N5 carryovers are intentionally absent. Exam mechanics (sections, scoring, sitting dates) live in The JLPT Explained: Levels, Sections, and What Each Means, and general word-count and coverage theory lives in How Many Japanese Words Do You Need to Be Fluent?; neither is re-derived here.
How N1 builds on N2
N1 (cumulative ~10,000) is approximately the full N2 base (~6,000, which already contains the N3, N4, and N5 layers) plus roughly 2,700–3,300 new words.45678 This page focuses on that new layer. The JLPT N2 Vocabulary List holds the prerequisite base, and the JLPT N3 Vocabulary List, JLPT N4 Vocabulary List, and JLPT N5 Vocabulary List hold the foundation under it.
That ~2,700–3,300 band comes from direct counts of new-words-only lists, not from subtraction. The Tanos new-words-only N1 list counts 2,698 entries;56 jpdb's internal new-at-N1 delta is 3,332;7 the Amenokori N1 deck is 3,239.8
Computing new-at-N1 as cumulative N1 (~10,000) minus cumulative N2 (~6,000) gives ~4,000. That overstates the real layer by roughly 700–1,300 words against every direct count above. The cumulative round figures and the new-words lists are built on different bases and at different scopes, so the gap between two coarse estimates is not a reliable count of the delta. Use the direct-counted ~2,700–3,300 band, never the ~4,000 subtraction.
The shift from N2 to N1 is qualitative, not just larger. This is where Japanese moves from "functional in business and news" into literary, abstract, specialized, and written-only registers. N2 reading covers newspaper and magazine articles and commentaries, while N1 reading covers editorials, critiques, and abstract writings with logical complexity.2
Why N1 is read, not drilled: the long-tail shift
Unlike N5 through N2, where a finite categorized list is genuinely drillable, the new-at-N1 layer is a long tail of low-frequency vocabulary. It contains thousands of literary, formal-written, and specialized words that each appear rarely. Word-frequency distributions in Japanese written text are heavily long-tailed, so a small high-frequency core covers most running text and coverage gains per additional word fall off sharply once you are past that core.9
The new-at-N1 words sit out in that tail, which is why flashcard return-on-investment drops sharply for them. The exam-strategy companion JLPT N1 Vocabulary: The Long-Tail Problem treats this diminishing-returns problem from the test-prep angle. The JLPT-specific coverage curve is not published, so this rests on the corpus-scale frequency principle rather than an N1-specific number.9
The practical strategy is to read widely (novels, editorials, non-fiction, academic text) and sentence-mine the unknown words you actually meet. Prioritize words by personal frequency rather than by a list's order.9 The Sentence Mining article covers the mining workflow, and Word-Frequency Coverage covers the why behind the flattening.
Most of the N1 long tail only needs to be passive, meaning recognized rather than produced. This fits the exam, which tests recognition. Active vs. Passive Vocabulary explains how that split changes the way you study these words.
Why "~10,000 cumulative" and not an exact new-layer number
The count is unofficial community consensus, not a JEES list.3 The cumulative N1 total clusters around ~10,000 across sources (Tanos ~10,000;4 JLPTsensei "about 10,000"1) because each source reconstructs the post-2010 scope differently. The new-layer delta is itself a range: the Tanos new-words-only list is 2,698,56 jpdb's internal delta is 3,332,7 and the Amenokori N1 deck is 3,239,8 so "~2,700–3,300 new" is the honest direct-counted band. Sources differ by hundreds to thousands of words because each reconstructs the list on a different basis. The long tail also has no hard boundary.
How to read this list
The tables below give the kanji form, kana reading, optional romaji, English gloss, and a register or domain note for each entry. This page shows words new at N1, so common N2/N3/N4/N5 carryovers are intentionally absent. The cumulative N1 inventory is the N2 list plus these.5
Many entries are written-register or literary and rare in speech. Each table is a representative sample of its register, not an exhaustive four-figure dump. As the read-don't-drill section above explains, the long tail makes exhaustiveness both impossible and useless.
What's new at N1: vocabulary by register
The core of an N1 list is the delta, organized by the registers N1 adds rather than presented as one flat list. Each group below is a compact table of words new at N1, confirmed on the Tanos new-words-only N1 list and cross-checkable against JLPTsensei.561 These are representative anchors, not the full ~2,700–3,300-word inventory.
Example sentences are either verbatim Tatoeba sentences cited by ID, or labeled constructed minimal examples where no clean corpus sentence was available for a low-frequency literary head word.
Literary and written-only register (文語・書き言葉)
This is the literary, classical-flavored, and written-only layer that separates N1 from N2. It includes formal connectives, written-register words, literary verbs and adjectives, and words that fill novels and editorials but rarely surface in speech. You meet this register by reading fiction and editorials. It is where the long tail begins. The classical grammar these words sit inside is covered in Bungo (文語) Grammar Primer for Modern Readers.
| Kanji form | Reading | Romaji | Gloss | Register note |
|---|---|---|---|---|
| 最早 | もはや | mohaya | (no) longer, already, by now | written / formal connective |
| 凡そ | およそ | oyoso | roughly, generally, in general | written-register adverb |
| 専ら | もっぱら | moppara | wholly, solely, entirely | written / formal adverb |
| 敢えて | あえて | aete | dare to, venture to, go so far as to | written / formal adverb |
| 悉く | ことごとく | kotogotoku | entirely, altogether, one and all | literary adverb |
| 甚だ | はなはだ | hanahada | exceedingly, greatly | literary intensifier |
| 強いて | しいて | shiite | by force, against one's will | written-register adverb |
| 顧みる | かえりみる | kaerimiru | to look back on, to reflect on, to review | written / literary verb |
| 赴く | おもむく | omomuku | to proceed to, to go toward, to head for | written / formal verb |
| 阻む | はばむ | habamu | to hinder, to obstruct, to check | written / formal verb |
The two examples below are constructed minimal sentences, not corpus-sourced. No clean verbatim Tatoeba sentence was available that isolated these low-frequency head words. The vocabulary itself is confirmed new at N1.56
最早手遅れだ。
"It is too late now."
強いて言えば、こちらの方が好きだ。
"If I had to choose, I prefer this one."
These are written-register words. They appear in novels, editorials, and formal prose but are rare in casual speech, so learners usually meet them by reading fiction and editorials, not through conversation.25
Formal-written Sino-Japanese 漢語 (硬い漢語)
This is the densest 漢語 (Sino-Japanese vocabulary) layer. It consists of abstract, formal, low-frequency compounds of two or more kanji that pattern academic and official writing. Many also double as する-verbs. It continues the 漢語 surge that began at N2, but with heavier and rarer words. The Jukugo (熟語): How Kanji Combine to Form Japanese Words article covers how these compounds are built and read, and Suru-Verbs (する-Verbs) covers their verbalized forms.
| Kanji form | Reading | Romaji | Gloss | Also する-verb? |
|---|---|---|---|---|
| 概念 | がいねん | gainen | concept, general idea, notion | n/a |
| 妥協 | だきょう | dakyō | compromise, giving in | yes (妥協する) |
| 該当 | がいとう | gaitō | corresponding to, falling under | yes (該当する) |
| 矛盾 | むじゅん | mujun | contradiction, inconsistency | yes (矛盾する) |
| 是正 | ぜせい | zesei | correction, rectification | yes (是正する) |
| 措置 | そち | sochi | measure, step (taken) | yes (措置する) |
| 趣旨 | しゅし | shushi | purport, gist, object, point | n/a |
| 崇拝 | すうはい | sūhai | worship, adoration, reverence | yes (崇拝する) |
ゼロの概念はインドで発明されました。10
"The concept of 'zero' originated in India."
彼は矛盾している。10
"He isn't consistent with himself."
この条件に該当する人は誰もいない。10
"There is nobody who fulfils these conditions."
ついに父は妥協した。10
"Finally, my father compromised."
Kango makes up roughly 60% of dictionary vocabulary but only about 18–20% of ordinary speech. Its share rises sharply in formal, abstract, and academic register,11 which is exactly the register this bucket adds. The proportion of kango among new-at-N1 words is therefore high and skews written.
Specialized and academic vocabulary (専門・学術)
This is the domain and technical layer N1 reading draws on: academic, legal, economic, scientific, and field-specific terms that appear in non-fiction and specialist text. It is inherently long-tail, because you cannot pre-learn every field's terms. Instead, you mine the fields you actually read.
| Kanji form | Reading | Romaji | Gloss | Domain |
|---|---|---|---|---|
| 細胞 | さいぼう | saibō | cell (biology) | science |
| 分子 | ぶんし | bunshi | molecule; numerator | science / math |
| 繁殖 | はんしょく | hanshoku | breeding, propagation | biology |
| 融資 | ゆうし | yūshi | financing, loan | finance |
| 株式 | かぶしき | kabushiki | stock, shares | finance |
| 赤字 | あかじ | akaji | deficit, being in the red | economics |
| 司法 | しほう | shihō | administration of justice, judiciary | law |
| 弁護 | べんご | bengo | defense, pleading, advocacy | law |
| 官僚 | かんりょう | kanryō | bureaucrat, bureaucracy | politics / government |
| 世論 | せろん / よろん | seron / yoron | public opinion | politics / media |
| 抗議 | こうぎ | kōgi | protest, objection | civics / news |
トムは銀行から多額の融資を受けた。10
"Tom borrowed a lot of money from the bank."
私はその抗議を退けた。10
"I dismissed the protest."
This bucket is field-dependent. A finance reader meets 融資, 株式, and 赤字 constantly, while a biology reader meets 細胞, 分子, and 繁殖. Neither pre-learns the other's terms. Acquisition is routed to the read-and-mine strategy below.
Idiomatic expressions: 慣用句 and 四字熟語 at N1
N1 reading and listening lean on idioms, set phrases, and four-character compounds whose meaning is non-compositional, meaning you cannot reliably infer it from the parts. This section names the category and representative N1 members. The idiom cluster explains the mechanics. For drilling the category, see Yojijukugo: Reading and Using Four-Character Idioms, Kotowaza: Japanese Proverbs, Japanese Body-Part Idioms: 手, 目, 口, 心 Expressions, Animal Idioms: 馬, 犬, 猫, 虎, and the kanji-side Yojijukugo (四字熟語): The Japanese Four-Character Idioms Explained.
| Kanji form | Reading | Romaji | Gloss | Note |
|---|---|---|---|---|
| 相次ぐ | あいつぐ | aitsugu | to occur one after another, to follow in succession | set verb, written / news |
| 相応 | そうおう | sōō | suitability, befitting, in keeping with | 漢語 set phrase |
| 相応しい | ふさわしい | fusawashii | appropriate, befitting, suitable | set adjective |
| 余儀ない | よぎない | yoginai | unavoidable, inevitable | set adjective, formal |
| 不審 | ふしん | fushin | suspicious, dubious, doubt | 漢語 set noun |
Specific four-character idioms are deferred to the yojijukugo articles by title rather than asserted as confirmed N1 here. The reason is that the Tanos N1 list enumerates words, not a clean 四字熟語 sublist.5
新型コロナウイルスの感染拡大防止のために、ライブイベントの自粛が相次ぐ中、10
"Amid a succession of self-imposed restraints on live events to help prevent the spread of the coronavirus,"
The low-frequency long tail
Beyond the representative buckets above lies the bulk of the new-at-N1 layer: thousands of words that each appear rarely. This is where flashcard return-on-investment is lowest and personal-frequency mining works best. This follows from the heavy long tail in Japanese written-text frequency distributions. Once you are past the high-frequency core, each additional word adds less coverage.9
| Kanji form | Reading | Romaji | Gloss |
|---|---|---|---|
| 嗚呼 | ああ | ā | Ah!, Oh!, Alas! (literary interjection) |
| 山岳 | さんがく | sangaku | mountains, alpine |
| 座標 | ざひょう | zahyō | coordinates |
| 寒気 | さむけ | samuke | chill, the shivers |
| 酸 | さん | san | acid |
| 悟る | さとる | satoru | to attain enlightenment, to perceive, to realize |
This bucket is the honesty payload of an N1 list. It is where a flat list stops being the right tool. Each word here is too rare to justify deck space yet matters when you personally meet it. The move is reading widely and sentence-mining, as the read-don't-drill section, Sentence Mining, and Word-Frequency Coverage lay out. N1 is not finishable as a list.
N1 kanji coverage
N1 vocabulary is written almost entirely in kanji and assumes essentially the full jōyō set. The current jōyō list comprises 2,136 characters (2010 revision).12 Community reconstructions put the N1 kanji requirement at about 2,000 characters cumulative, with roughly 1,504 treated as the N1-specific layer.13 Tanos likewise summarizes the N1 kanji requirement as ~2,000.4
These N1 kanji figures are themselves unofficial, since no official list has existed since 2010.3 The safe statement is that N1 assumes essentially the whole jōyō set.12 The formal and literary register also uses less-common readings of common characters and, on occasion, beyond-jōyō characters. N1 reading is therefore not strictly bounded by the jōyō list.
For the drill targets and the realistic count, see Secondary School Jōyō Kanji (中学校 + 高等学校): The 1,110-Character Set Beyond Elementary, The Jōyō Kanji List (常用漢字): The 2,136-Character Set Explained, and How Many Kanji Do You Need? A Realistic Count.
The N1 layer is so 漢語-heavy because Sino-Japanese vocabulary dominates dictionary entries and becomes more common in formal and abstract register.11 Wago, Kango, Gairaigo, Konshugo: The Four Vocabulary Strata of Japanese explains why. This is also why kanji-and-vocab co-study pays compounding interest at N1. The same on'yomi recurs across many compounds, and many N1 nouns double as する-verbs.
How to actually learn the new N1 words
The N1-specific method follows directly from the long-tail reality and the broad written register the exam targets.29 Read widely (novels, editorials, non-fiction, news, academic text). Sentence-mine unknowns into an SRS, prioritize by personal frequency, and accept that much of the long tail stays passive. General acquisition theory lives in How to Learn Japanese Vocabulary: A Strategy by Level; the workflow specifics live in Sentence Mining, Word-Frequency Coverage, and Active vs. Passive Vocabulary. For parsing the editorial, literary, and academic prose where these words live, see JLPT N1 Reading: Literary, Academic, and Editorial Texts.
Be honest about the contrast with lower levels. A fixed pre-made N1 deck covers the high-frequency slice well. The long tail is best mined from your own reading, because each tail word is too rare to justify deck space yet matters when you personally meet it.9 Kanji-and-vocab co-study still pays compounding interest given the 漢語 density.
Recommended deck and list sources
The community-standard N1 sources are the basis for the figures in this article. They cover the high-frequency slice well, while the long tail should be mined from reading.
- Tanos N1 vocabulary list: the de-facto community new-words-only N1 list, explicitly not cumulative.45
- jpdb N1: frequency-ordered and corpus/media-derived, the best raw acquisition source and a direct count of the layer.7
- JLPTsensei N1 vocabulary list: a paginated reference with readings and glosses.1
- Commercial series: JLPT Tango N1, Sou-Matome N1, and Shin Kanzen Master N1 are commonly cited community-standard study series.
Amenokori app callout
No deck reaches the whole N1 tail. For the part that pays back fastest, J-Compass recommends Amenokori: its N1 set drills the high-frequency N1 vocabulary (3,239 words plus an 803-word extended set) on an FSRS schedule. Your own reading and mining handle the long tail around it.8 For how it fits the wider toolkit, see Choosing Your First Japanese Resources: Free vs. Paid and How to Learn Japanese Vocabulary: A Strategy by Level.
Good to know
The N1 long tail is written-register, not spoken vocabulary
Most new-at-N1 words appear in editorials, novels, and academic prose and are rare in casual speech. Examples include the literary adverbs 専ら and 凡そ, formal 漢語 compounds, and specialist terms.25 Pairing the list with heavy reading immersion is the highest-leverage move at this level, because that is where these words actually occur.
Treating the cumulative ~10,000 as the new study target
A learner who decides "I must drill ~10,000 N1 words," or "~4,000 new words" by subtracting cumulative N2, is chasing the wrong number. The ~10,000 is the cumulative total and includes the entire N2/N3/N4/N5 base. The naive subtraction overstates the new layer.56 Drill the ~2,700–3,300 high-frequency new layer (direct-counted 5678) and maintain the N2 base. Then read and mine the rest. The exam still draws on the whole base, so do not neglect it while chasing the new layer.3
The exam tests recognition, so most of the tail only needs to be passive
N1 measures reading and listening comprehension, not production.2 A learner can therefore pass while only recognizing most tail words rather than producing them. That changes how you study them, which Active vs. Passive Vocabulary treats directly.
The 漢語 density means kanji pays compounding interest
Sino-Japanese vocabulary dominates formal and abstract register.11 Learning one character's on'yomi therefore unlocks many N1 compounds at once. Many N1 nouns also double as する-verbs, which stretches each entry further across reading and listening.
There is no single "correct" 10,000-word list
Lists differ by hundreds to thousands of words because each reconstructs the post-2010 scope differently: by exam-spec mapping, corpus frequency, or deck curation. The long tail has no hard boundary.357 The general phenomenon is frequency flattening across the tail, which Word-Frequency Coverage treats rather than this page.
See also
- JLPT Vocabulary by Level: How Many Words for N5 to N1
- JLPT N3 Vocabulary List: ~1,800 New Words Beyond N4, by Domain
- JLPT N4 Vocabulary List: ~700 New Words Beyond N5, by Category
- JLPT N5 Vocabulary List: ~800 Words by Category, Kanji Coverage, and Decks
- Secondary School Jōyō Kanji (中学校 + 高等学校): The 1,110-Character Set Beyond Elementary
- Yojijukugo (四字熟語): The Japanese Four-Character Idioms Explained
- Choosing Your First Japanese Resources: Free vs. Paid