Reading (読解) never gets a clock of its own. At N5, N4 and N3 it is bolted onto the grammar block. At N2 and N1 it shares a single block with grammar and vocabulary together, and from N3 upward it also carries a score floor you have to clear on its own marks. That is where the section is actually lost, and the problem is structural rather than linguistic. What changes as you climb is which of the six official question types can turn up on your paper, and how much of a passage you have to hold in your head at once.
Reading never gets its own clock
Which block reading lands in decides how much of the clock is really yours, and the answer changes twice on the way up.
| Level | The block reading sits in | Vocabulary | Listening | Total |
|---|---|---|---|---|
| N5 | Grammar・Reading, 40 min | separate, 20 min | 30 min | 90 min |
| N4 | Grammar・Reading, 55 min | separate, 25 min | 35 min | 115 min |
| N3 | Grammar・Reading, 70 min | separate, 30 min | 40 min | 140 min |
| N2 | Vocabulary/Grammar・Reading, 105 min | inside the block | 50 min | 155 min |
| N1 | Vocabulary/Grammar・Reading, 110 min | inside the block | 55 min | 165 min |
Look at what happens between N3 and N2. At N5, N4 and N3, vocabulary is fenced off in a block of its own, so it cannot eat your reading time. Only grammar can. At N2 that fence comes down, and vocabulary, grammar and every passage share one clock. It is why people describe N2 as the level where they ran out of time even though the passages were barely longer than N3's.
How reading is scored, and why N3 is the cliff
The JLPT does not simply add your marks up. It scores you by section and demands a minimum in each one, on top of the overall total. Reading's status in that scheme changes at N3.
| Level | Is Reading its own scoring section? | Score range | Sectional floor | Overall pass mark |
|---|---|---|---|---|
| N1 | Yes | 0 to 60 | 19 | 100 / 180 |
| N2 | Yes | 0 to 60 | 19 | 90 / 180 |
| N3 | Yes | 0 to 60 | 19 | 95 / 180 |
| N4 | No, merged with Language Knowledge | 0 to 120 | 38 | 90 / 180 |
| N5 | No, merged with Language Knowledge | 0 to 120 | 38 | 80 / 180 |
N3 is where reading starts carrying its own score. Below it, a strong grammar and vocabulary result literally adds to a weak reading result, because they land in the same 0 to 120 bucket. From N3 up they do not: reading needs 19 of its own 60, and nothing else on the paper can cover for it.
The six question types, and which levels use them
The JLPT publishes exactly six reading item types. They are not spread evenly, and knowing which ones can appear on your paper tells you what to practise.
| Reading item type | N5 | N4 | N3 | N2 | N1 |
|---|---|---|---|---|---|
| Comprehension, short passages (内容理解・短文) | ● | ● | ● | ● | ● |
| Comprehension, mid-size passages (内容理解・中文) | ● | ● | ● | ● | ● |
| Comprehension, long passages (内容理解・長文) | ○ | ○ | ● | ○ | ● |
| Integrated comprehension (統合理解) | ○ | ○ | ○ | ● | ● |
| Thematic comprehension, long (主張理解・長文) | ○ | ○ | ○ | ● | ● |
| Information retrieval (情報検索) | ● | ● | ● | ● | ● |
Two rows are worth staring at. Comparative reading is an N2 invention. 統合理解 puts two short texts on the same topic in front of you and asks where they agree, which is a working-memory task before it is a reading one, and nothing at N3 prepares you for it. Neither text is hard on its own. That is the trick.
Aの筆者は、在宅勤務は通勤時間をなくすので生産性が上がると述べている。
直訳 Word by word
A's writer, working from home, commuting time because it removes, productivity rises, states.
Text A, stated flatly. Every word here is N3 vocabulary.
一方、Bの筆者は、同僚と偶然話す機会が減る点を問題にしている。
直訳 Word by word
On the other hand, B's writer, with colleagues by chance talking chances decrease, that point is treating as a problem.
Text B, also easy. Now answer what A and B have in common. Neither text says it, and that gap is the entire question.
Reading text A, forming an answer, then reading text B to confirm it. Every option in 統合理解 is written to be true of one text or the other, so an option that faithfully restates A is still wrong when B never said it. Read both, state each writer's position in one clause, and only then look at the four options.
The other row is subtler. At N2, the only long passage on the paper is 主張理解. N1 sets both kinds of long text, N2 sets the thematic one alone. So where N3 asks what the passage said, N2 asks what the writer was arguing, and every hour of N3 practice leaves that muscle untrained.
JLPT N5 reading: sentences, not passages
The official can-do statement for N5 reading is modest and accurate: "One is able to read and understand typical expressions and sentences written in hiragana, katakana, and basic kanji."
There is no long passage at N5 and no inference. You get a short text (a note, a memo, a message), a mid-size one (a paragraph about a familiar situation), and information retrieval (find one fact in a poster or a timetable). Every answer sits on the surface of the text.
あしたの午後三時に、二階の教室に来てください。
直訳 Word by word
Tomorrow's afternoon three o'clock at, second floor's classroom to, please come.
All of N5 reading is in this line. One fact, sitting on the surface, and a question that points straight at it.
What costs N5 candidates marks is decoding speed. They read kana slowly enough that the 40-minute block runs out while they are still sounding words out, and no reading strategy fixes that. Kana fluency does. Work through hiragana and katakana until recognition is automatic, then start ticking off the JLPT N5 kanji and vocabulary list.
JLPT N4 reading: the same shapes, bigger
N4 uses the same three item types as N5, and the official can-do only widens a little: "One is able to read and understand passages on familiar daily topics written in basic vocabulary and kanji."
What grows is the text. The mid-size passage becomes a real paragraph with a shape: a small narrative, or an explanation with a reason buried inside it. Information retrieval starts using documents that fight back, like a leaflet with conditions attached or a schedule with an exception printed under it.
会議は水曜日の予定でしたが、部長の出張が延びたため、金曜日に変わりました。
直訳 Word by word
The meeting, Wednesday's plan was, but, the section chief's business trip extended because, to Friday it changed.
The question asks which day the meeting is. 水曜日 is the first day named, it sits right next to the keyword, and it is wrong.
Answering from the first sentence that contains the question's keyword. N4 is where the test starts punishing this on purpose: the qualifier that flips the answer is almost always in the final clause of the paragraph. Read to the full stop before you look at the options.
N4 is the last level where reading cannot fail you by itself, and the right place to spend that safety on stamina: texts slightly too long for you, read without a dictionary. Keep the N4 kanji and vocabulary list moving underneath, because N3 will not wait for it.
JLPT N3 reading: the long passage arrives
N3 is the structural break in this progression. Three things land at once: the long content passage (内容理解・長文) appears for the first time, reading becomes its own scoring section with its own 19-point floor, and the grammar-plus-reading block stretches to 70 minutes without stretching far enough.
What N3 really introduces is text that will not state its point flatly. Opinion passages quote a view before disagreeing with it, and questions start asking what the author implies. The concession-then-contrast pattern (確かに…しかし…) is the highest-yield thing to be able to spot, because the sentence before しかし is where the wrong answer comes from.
確かに、紙の辞書には紙の辞書のよさがある。しかし、調べる回数が増えるほど、読む速さは落ちていく。
直訳 Word by word
Certainly, in paper dictionaries there is a paper dictionary's goodness. However, the more the number of look-ups increases, reading speed falls.
確かに sets up a view the writer is about to drop, and the distractor is built out of that first sentence. The answer lives on the other side of しかし.
Every tactic for that (questions before passage, connector tracking, time-boxing the four passage types, and the 30-day plan to build the speed it all depends on) is worked through in the JLPT N3 reading comprehension strategies guide. If N3 is your exam, go there next, and keep the N3 kanji and vocabulary list ticking over while you do.
JLPT N2 reading: two texts, one question
At N2 the passages stop being about things and start arguing things. The official can-do names the material outright: "articles and commentaries in newspapers and magazines as well as simple critiques."
Two new item types arrive and the fences come down. 統合理解 hands you two texts and asks you to compare them. 主張理解 hands you a long argumentative essay and asks what the writer is claiming. Both reward a reader who is tracking who is speaking: the hedges that mark a cited view, against the markers that mark the writer's own.
日本の労働時間は減少しているとされている。しかし、それは統計上の話に過ぎないのではないか。
直訳 Word by word
Japan's working hours are decreasing, it is said. However, that is a matter of statistics only, is it not?
Two voices in two sentences. 〜とされている hands you somebody else's claim. 〜のではないか is the writer stepping forward. 主張理解 asks about the second one.
Meanwhile vocabulary, grammar and reading now share one 105-minute clock, so every minute overspent on a grammar question is a minute stolen from a passage.
確認At N2 you finish the vocabulary and grammar questions with 40 minutes left on the clock. How much reading time do you have?
Forty minutes. There is no reading timer at any level, and at N2 the vocabulary, grammar and reading questions all sit inside one 105-minute block. Whatever you overspend early, you have taken off a passage at the end.
The full toolkit is in the JLPT N2 reading strategies guide: the T-chart for comparative passages, the hedging tracker, and a realistic 105-minute budget. The N2 kanji and vocabulary list is what makes any of it survivable at speed.
JLPT N1 reading: everything, at once
N1 is the only level that uses all six item types, and the only one carrying both kinds of long passage: 内容理解・長文, where the question is what the text says, and 主張理解・長文, where the question is what the writer is doing with it. The official can-do is explicit about the material: "writings with logical complexity and/or abstract writings on a variety of topics, such as newspaper editorials and critiques."
The overall pass mark rises to 100 out of 180, the highest of the five levels, while reading's own floor stays at 19. At N1 you cannot scrape through on section minimums. You need real marks, out of a 110-minute block that also holds every vocabulary and grammar question on the paper.
What makes N1 reading hard is rarely the vocabulary and almost never the grammar. It is that the argument is distributed: a claim is set up in paragraph one, qualified in paragraph three, landed in paragraph five, and the question asks about the version in paragraph five.
Answering a 主張理解 question from the paragraph where the claim first appears. At N1 the opening statement of a thesis is usually the version the writer is still willing to modify, and the option built from it will look right in every particular except the one being tested. Find the last place the writer states the claim, not the first.
Holding an argument across five paragraphs is a habit you build on editorials, not on practice papers. Read 天声人語 or an opinion column daily, and make a habit of stating the writer's thesis in one sentence when you finish. If you have never read native Japanese content outside a textbook, start there rather than on a mock paper. Keep the N1 kanji and vocabulary list ticking over underneath, because N1 vocabulary is specialised by domain and only breadth fixes that.
What every level rewards
Two habits transfer the whole way from N5 to N1, and neither costs you anything.
Read the question before the passage. Every level hands you the question and all four options up front, so reading them first converts an open comprehension task into a search with a target. You stop reading in order to understand, and start reading in order to find, which is a faster and much less tiring activity.
Then treat connectors as structure rather than vocabulary. しかし, つまり, 確かに and 一方 are not words to translate. They are the writer telling you, for free, where the answer is about to arrive, and they behave the same way in an N3 opinion paragraph and an N1 editorial.
一方、地方では人口が減り続けている。つまり、同じ政策が同じ結果を生むとは限らない。
直訳 Word by word
On the other hand, in the regions the population continues to decrease. In other words, the same policy the same result produces, is not necessarily so.
一方 announces a contrast. つまり announces that the writer is about to say the thing plainly. Two instructions about where to look, before you have understood a single clause.
Everything past those two is level-specific, which is why the depth lives on the N3 and N2 pages rather than here. What is not level-specific is the mileage. Strategies only fire at reading speed, and reading speed comes from reading.
The list nobody publishes
One last thing, because it changes how you should read every kanji count you have ever seen for this exam. The JLPT does not publish a kanji list, a vocabulary list, or a grammar list. It stopped after the 2010 revision and said why: the organisers decided that publishing "Test Content Specifications" containing a list of vocabulary, kanji and grammar items "was not necessarily appropriate" for a test that measures communicative competence rather than memorisation.
So every "N3 = 650 kanji" figure in circulation, including the ones on our own level indexes, is a reconstruction, assembled from the pre-2010 specification and from what actually turned up in past papers. Aim at them by all means, but expect the paper to reach outside them.
Which makes which kanji can you read right now the far more useful question. Work through the grid below and it will estimate the level you are reading at, so you can pick the strategy guide that matches the exam you are actually sitting.