How Many Kanji Do You Need for Each JLPT Level?

“How many kanji do I need for N3?” is one of the most-asked questions in Japanese learning, and the internet answers it with a strange mix of precision and hedging. Here’s the straight version: the numbers, where they come from, why they’re not official anymore, and — more importantly — what “knowing” a kanji for the JLPT actually means.

New to the exam itself? The JLPT N5 guide explains how the test works from the first level up.

The numbers: kanji per JLPT level

Three things to read off that table:

• The counts are cumulative. N3’s ~650 includes the N5 and N4 sets. You’re always building on the same pile.
• The jumps get bigger. N4→N3 roughly doubles your kanji; N3→N2 adds ~350 more; N2→N1 adds another ~1,000. This is a big part of why the N2→N1 gap is famously the hardest in the whole exam series.
• N1 ≈ the jōyō set. Japan’s official list of 2,136 jōyō (“everyday use”) kanji is what schools teach through high school. N1 covers essentially that territory — pass it and you recognize roughly what a Japanese 18-year-old does.

Why these numbers are “unofficial” (and why they’re still right)

Until 2010, the JLPT published exact kanji and vocabulary lists. The 2010 revision removed them on purpose: the Japan Foundation wanted to test whether you can use Japanese, not whether you memorized a checklist.

In practice, the exam still draws from the same frequency-based pools it always did — everyday-frequency kanji at lower levels, newspaper- and literature-frequency at the top. Community lists compiled from past exams have tracked actual test content closely for over a decade. So: use the numbers as solid planning targets, but expect the occasional out-of-scope character (lower levels typically add furigana when that happens — see what furigana is if that’s new to you).

Why raw counts flatter you: the readings problem

Here’s what “knowing 650 kanji” hides. Most kanji have multiple readings — a Chinese-derived on’yomi and one or more native kun’yomi — and the exam tests them in real words, where you can’t dodge:

• 生 appears in 学生 (gakusei, student), 生きる (ikiru, to live), 生まれる (umareru, to be born), and 生 (nama, raw) — four different readings of one “known” kanji
• 日 alone runs through 日曜日 (nichiyōbi), 今日 (kyō), 毎日 (mainichi), and 二日 (futsuka)

A kanji learned as an isolated flashcard with one English keyword is maybe a third of a kanji by exam standards. The on’yomi vs kun’yomi guide unpacks the system — the short version is that readings are learned reliably only one way: inside real words, met repeatedly in context.

The flip side is genuinely good news: the JLPT is 100% multiple choice. You never write a single kanji by hand at any level. Recognition is the whole game, and recognition builds much faster than production.

The payoff curve: why the first 1,000 matter most

Kanji study follows a steep frequency curve. Corpus analyses of Japanese newspapers consistently show that the most frequent 500 kanji cover roughly 80% of characters in running text, and the top 1,000 push past 90%. The second thousand — the N1 territory — buys you the long tail: precise, literary, technical, and formal vocabulary.

Practical consequences:

1. N5→N3 kanji transform your daily life in the language: menus, signs, messages, simple articles all start opening up.
2. N2 (~1,000) is the functional reading threshold — the level where native material becomes navigable with a dictionary rather than impenetrable.
3. N1’s extra thousand is about precision, not survival. It’s what employers and universities use as the bar because it signals you can handle any text, not just common ones.

So if the 2,000 figure feels crushing: the curve is on your side. Every kanji you learn early pays out constantly; the expensive ones at the end are the cheapest to contextualize because by then you read all the time.

How to actually learn them (the method beats the count)

The counts tell you how many; they say nothing about how. Decades of learner experience converge on the same points — the full method is in the kanji learning guide, but the essentials:

• Learn kanji through words, not in isolation. 食 as a flashcard is trivia; 食べる, 食事, 昼食 are vocabulary with readings built in. Two birds, every stone. The vocabulary guide covers this side.
• Use spaced repetition for retention, reading for consolidation. SRS gets a kanji into memory; meeting it in real sentences is what makes it stay and speeds up recognition.
• Pace for the long haul. 5–10 new kanji a day, every day, beats 50-kanji weekend binges — the forgetting curve punishes bursts.
• Read at your level from the start. This is where most plans fail: native material is too hard for years, so learners never get the reading volume that makes kanji automatic. Level-graded reading fixes it — how to read Japanese lays out the progression.

Where Shinobi fits

Kanji counts are climbed by reading, and reading only works when the text matches your level. Shinobi’s graded stories are organized exactly along the JLPT scale — pre-N5, N5, N4, N3 and up — so the kanji in your stories are the kanji on your next exam, met in real sentences with furigana, native audio, and tap-to-translate. Instead of grinding a 650-character list for N3, you read stories where those 650 characters keep showing up until recognizing them is automatic. Browse the full story library to start at your level.

The bottom line

~100 for N5, ~300 for N4, ~650 for N3, ~1,000 for N2, ~2,000+ for N1 — cumulative, unofficial since 2010, and still the right planning numbers. But the count is the map, not the journey: the JLPT tests kanji as they live inside words and passages, so the learners who pass are the ones who read daily at their level, not the ones with the biggest flashcard decks.

For the exam-by-exam picture — format, sections, pass marks — start with the JLPT N5 guide, and see how long the whole journey takes to plan your timeline.