Short answer: if you already passed N3, budget 6 months at 8–10 hours a week, or 3 months at 15–20 hours a week. Spend roughly 40% of that time on vocabulary and kanji, 20% on grammar, 25% on timed reading, and 15% on listening. The single biggest reason people fail N2 after passing N3 is not grammar — it's that they never practise reading under the clock, and they run out of time on the reading section.
I'm studying for N2 myself right now, and I'm building betterlingo.io while I do it, so this is the plan I actually follow rather than a plan I invented for a blog post. Take the numbers as a starting shape, not gospel.
What N2 actually asks of you
The exam is 180 points total, and you pass at 90/180 — but there's a trap in that number. There are three scoring sections:
| Section | Points | Sectional minimum |
|---|---|---|
| Language knowledge (vocab + grammar) | 60 | 19 |
| Reading | 60 | 19 |
| Listening | 60 | 19 |
You must clear both the total (90) and every sectional minimum (19 each). A 120 overall with an 18 in listening is a fail. This matters more than it sounds, because the natural way to study — grinding vocab decks because they're easy to measure — inflates exactly one section and starves the two that are hard to fake.
Timing: 105 minutes for language knowledge + reading combined, then 50 minutes for listening. That combined block is the crunch. There's no per-part timer, so it's entirely on you not to spend 50 minutes on vocab and grammar and arrive at the long passages with 20 minutes left.
Rough scale of the content: about 6,000 words and 1,000 kanji, against N3's ~3,700 words and ~650 kanji. So N3 → N2 is not an incremental step; it's most of another N3 stacked on top.
The 6-month plan (8–10 hrs/week)
This is the version I'd recommend to almost everyone. Assumes you're at solid N3.
Months 1–2: volume
- Vocabulary/kanji, ~30–40 min/day. New cards at a number you can sustain forever, not a number that feels heroic in week one. (Anki review load stabilises at roughly 8–10× your daily new count. Pick 15 new/day and you're signing up for ~150 reviews/day, every day, indefinitely. That's the actual contract.)
- Grammar, 2 sessions/week. Work through one N2 grammar book in order — Shin Kanzen Master 文法 or Sou Matome, either is fine. Don't shop for the perfect book; finish one.
- Reading, 1 session/week, untimed. Just build tolerance for length.
Months 3–4: production and timing
- Keep the vocab habit. It never stops.
- Start timed reading: one full 読解 set per week, under real time pressure, no dictionary.
- Add listening properly — one N2 listening set per week, plus passive input you actually enjoy the rest of the time. Passive input alone will not get you past 19/60; it builds the ear, it doesn't build the test skill.
- Start producing output. Write short answers in Japanese from memory instead of only recognising the right choice. More on this below.
Months 5–6: exam shape
- One full mock exam every 2 weeks, timed end to end, in one sitting. This is the part everyone skips and it's the part that decides the result.
- Every mock: mark it, then go back and reread the passages you got wrong untimed. The question you want to answer isn't "what was the right answer" — it's "where exactly did I lose the thread?" Usually it's one connective or one referent, and it's the same class of thing every time.
- Last 2 weeks: no new material. Consolidate, sleep, re-drill your weak section only.
The 3-month plan (15–20 hrs/week)
Same sequence, compressed, with two changes:
- Start timed reading in week 2, not month 3. You don't have the luxury of a slow ramp.
- Cut breadth, not reps. One vocab source, one grammar book, one listening source. In a 3-month window, switching resources is the most expensive mistake available to you — every switch costs you the review backlog you'd already built.
Realistically this works if you passed N3 comfortably and you have consistent daily time. If your N3 was a squeaker, take the 6 months. The exam runs twice a year (July and December); failing costs you six months, and the compressed plan is exactly the one that produces a 17 in listening.
The weekly template I actually use
| Day | Block |
|---|---|
| Mon | Vocab/kanji + grammar chapter |
| Tue | Vocab/kanji + listening set |
| Wed | Vocab/kanji + timed reading set |
| Thu | Vocab/kanji + grammar chapter |
| Fri | Vocab/kanji + free-form output |
| Sat | Longer session: mock section or review of the week's mistakes |
| Sun | Vocab/kanji only (light) |
Vocab every single day, everything else rotating. The rotation matters more than the length of any one block — an hour on Saturday does not replace 20 minutes on Wednesday.
The four skills, and what each one actually needs
文字・語彙 (vocab/kanji). Pure volume, and the most mechanisable part. SRS is genuinely the right tool here. One caveat: a card you've failed three or four times isn't a discipline problem, it's a bad card. Rewrite it with more context or delete it — leeches quietly eat a huge share of your review time and give nothing back.
文法 (grammar). N2 grammar is largely a set of fixed expressions with narrow usage conditions (〜わけではない、〜どころか、〜からといって…). Recognising them in a sentence is easy. Producing them is not, and the exam's 文の組み立て questions are much closer to production than recognition. If you can only ever pick the right one from four options, you're studying for a test you're not taking.
読解 (reading). This is a timed-volume problem, not a comprehension problem. Most people at N2 level can understand the passages if you give them unlimited time and a dictionary. The exam gives you neither. So the practice has to be: full sets, under the clock, dictionary closed, then a slow untimed reread of what you got wrong. Video courses and grammar explanations never put you under time pressure, which is why people who "studied a lot" still lose this section.
聴解 (listening). The one that fails people. Two things help disproportionately: (1) shadow the audio you've already understood, don't just re-listen passively; (2) practise the 即時応答 quick-response format specifically — it's a distinct skill from understanding a long conversation, and it's free points once the format stops surprising you.
The failure mode nobody warns you about
The most common shape of an N2 failure is: strong vocab score, passing grammar, reading in the low 20s, listening under 19.
It happens because multiple-choice study is comfortable. You can pattern-match your way through an MCQ deck without ever retrieving anything from scratch — you recognise the right answer rather than produce it, your accuracy looks great, and none of it transfers to a timed passage or a spoken prompt.
The fix is boring: make a meaningful part of your practice free-form. Answer a question in Japanese with nothing to choose from. Translate a sentence in both directions. Write one sentence using today's grammar point without looking at the pattern. It's much less pleasant than clicking options, which is exactly why it works — the effort is the point.
That's the itch I've been building BetterLingo to scratch, and I'll disclose it plainly: I'm the founder, and I'm also its heaviest user. Daily practice there is free-form rather than multiple choice — you type an answer, an LLM grades it against a rubric for that question type and shows you where it fell short, and there's a tutor that can see your actual attempt rather than answering in the abstract. It's rough in places and I fix things as I hit them in my own study. You do not need it to follow this plan; the plan works with a grammar book, an SRS deck, and a stack of past papers, which is how most people pass.
The 30-second version
- Passing N2 is 90/180 overall AND 19/60 in every section. Vocab grinding can't rescue a weak section.
- 6 months at 8–10 hrs/week from N3, or 3 months at 15–20 hrs/week if your N3 was comfortable.
- Split roughly 40% vocab/kanji, 20% grammar, 25% timed reading, 15% listening.
- Reading is a timed-volume problem — practise with the clock running, dictionary closed.
- Do free-form output, not only multiple choice, or you train recognition and get tested on recall.
- Full timed mocks from the halfway mark, every two weeks. Nothing else predicts your score as well.
If you're mid-N2 prep and doing something that contradicts the above, I'd genuinely like to hear it — the plan changes every time I learn something from a mock.
Top comments (0)