JLPT & JFT-Basic
Passing the JLPT takes more than the overall score
Passing the Japanese-Language Proficiency Test requires two things at once: an overall score at or above the pass mark, and a score at or above the minimum in every scoring section. One section below its minimum is a fail, however high the total. And JLPT scores are not a count of correct answers — they are scaled scores derived from your pattern of responses using item response theory, so counting "two more questions and I would have passed" does not match how the score is built.
Written by the United World Nihongo editorial team | Published 5 September 2026 | Sources: JLPT official site, "Scoring sections, pass/fail judgement and results"; the Japan Foundation and Japan Educational Exchanges and Services, "About scaled scores"; "Indication of the CEFR level for reference" | Confirmed on 5 September 2026 (確認日：2026年9月5日)
Two conditions, both required
The official site puts it this way, translated from the Japanese.
In order to pass, two things are necessary: (1) that the overall score is at or above the score required to pass (the pass mark), and (2) that the score in each scoring section is at or above the score required to pass set for that section (the section minimum).
JLPT official site, "Scoring sections, pass/fail judgement and results" (confirmed 5 September 2026)
| Level | Scoring sections | Range per section | Section minimum | Pass mark |
|---|---|---|---|---|
| N1 | Language knowledge / reading / listening | 0-60 each | 19 each | 100 |
| N2 | Language knowledge / reading / listening | 0-60 each | 19 each | 90 |
| N3 | Language knowledge / reading / listening | 0-60 each | 19 each | 95 |
| N4 | Language knowledge and reading / listening | 0-120 / 0-60 | 38 / 19 | 90 |
| N5 | Language knowledge and reading / listening | 0-120 / 0-60 | 38 / 19 | 80 |
Every level is out of 180. N1 to N3 have three sections of 60; N4 and N5 have one of 120 and one of 60. The totals come to the same number.
There is a usual shape to failing
Suppose you take N2 and score 40 on language knowledge, 33 on reading and 18 on listening. The overall score is 91, above the pass mark of 90. It is still a fail, because listening is one point short of the section minimum of 19.
Those numbers are constructed to show the mechanism; they are not from any statistic. The shape, though, is common. You can push the total up with reading and vocabulary and still fail on a listening score that stopped just short. Which means the studying that decides the result is not the studying that adds a few points to the total — it is the studying that lifts your weakest section over its minimum.
In this case no CEFR level is shown on the score document and an asterisk appears instead. The per-section scores are shown, so the result itself tells you where to work next.
The score is not a count of correct answers
This is the part most often misunderstood. A JLPT score is not a raw score adding up the marks for questions you answered correctly. It is a scaled score. The official explanation describes how it is produced.
The scaled score is determined mathematically by which questions were answered in which way — correctly or incorrectly — that is, by the "pattern of responses".
The Japan Foundation and Japan Educational Exchanges and Services, "About scaled scores" — translated from the Japanese (confirmed 5 September 2026)
The calculation rests on item response theory, and the same document states that it is entirely different from working out a raw score from the number of correct answers. Two consequences follow, both stated there.
- Two people with different numbers of correct answers, and different response patterns, can receive the same scaled score.
- Two people with the same number of correct answers can receive different scaled scores, because their response patterns differ.
The reason for the method is given as well. Keeping difficulty exactly constant from one sitting to the next is extremely difficult, so with raw scores the same ability would produce different scores depending on which test you happened to sit. Scaled scores quantify ability independently of that difficulty, so the same ability returns the same score whenever you take it.
One practical conclusion follows. "I was two questions short" is not a statement this test's scoring supports. Counting correct answers on past papers is useful for managing practice; it is not a reliable predictor of passing.
Why N3's pass mark is higher than N2's
In the table, N3 requires 95 and N2 requires 90 — the higher level asking for less. It looks like a mistake.
It is not, because comparing scores across levels was never part of what the scale claims to support. The scaled-score document is precise about the range within which comparison holds.
Where the level is the same and the scoring section is the same, the difference between scaled scores produced by two different sittings can be interpreted as a difference in Japanese-language ability, independently of the difficulty of the test.
The Japan Foundation and Japan Educational Exchanges and Services, "About scaled scores" — translated from the Japanese (confirmed 5 September 2026)
Same level, same section, different sittings. Putting N2's 90 beside N3's 95 and drawing a conclusion about which is harder falls outside what that sentence guarantees.
The same document explains how N3's pass mark was arrived at. The pass/fail standards of the current test were set to be consistent with those of the former test, so that someone able to pass the old Levels 1, 2, 3 and 4 can be understood as able to pass N1, N2, N4 and N5 respectively. N3 had no counterpart in the old test, so the pass/fail standards of the old Levels 2 and 3 were analysed statistically and N3's pass mark was set to fall between them.
In terms of ability, N3 sits between N2 and N4. That 95 is a larger number than 90 is an event on a different ruler.
What to do next
- Read the per-section scores, not just the verdict. Before the pass or fail, look at how far each section was from its minimum. The verdict is an outcome; the section scores are a plan.
- Start with the lowest section. Raising the total and clearing a minimum are different tasks. The second one decides the result.
- Measure yourself in situations, not in correct answers. Since the score comes from a response pattern, "how many I got right" is a weak indicator of where you actually are.
- Practise speaking and writing separately. The JLPT's measurement range is language knowledge, reading and listening; productive skills and interaction are not included. Passing does not mean those were assessed.
Frequently asked questions
I missed a section minimum by one point. Is there any allowance?
The conditions the official site gives are that the overall score reaches the pass mark and that each section reaches its minimum, and we found no statement of any exceptional treatment where a section falls short. The straightforward reading is that being one point short is a fail. For the next sitting, lifting that section above its minimum is the thing to work on first.
What score on past papers means I will pass?
You cannot estimate the result from correct answers on past papers. A JLPT score is not a raw total but a scaled score derived from the pattern of responses using item response theory, and the official explanation states that the same number of correct answers can produce different scaled scores where the response patterns differ. Past papers are useful for seeing which section is weak; they are not usable for predicting a score.
Can I decide between N3 and N2 by comparing the pass marks?
No, the numbers do not support that comparison. What the official scaled-score document describes as comparable is the same level and the same scoring section across different sittings. N3's pass mark of 95 and N2's of 90 belong to different levels, so setting them side by side says nothing about which is harder. Decide instead from the level of proof you actually need and from which of your sections is currently weakest.
If I pass, can I assume I can work in Japanese?
Check what the test measures. The JLPT official site states that its measurement range is language knowledge, reading and listening, and that productive skills — speaking and writing — and interaction are not included. Putting a view in a meeting, explaining a situation on the phone, writing an email: none of these is measured here. If that is the kind of situation giving you trouble at work, it needs practice separate from test preparation.