DEV Community

Cover image for N3's Pass Mark Is Higher Than N2's (And Other Quirks of JLPT Scoring)
Torres Wang
Torres Wang

Posted on Originally published at Medium

N3's Pass Mark Is Higher Than N2's (And Other Quirks of JLPT Scoring)

Why raw test scores lie, how the "hidden" section minimums wreck solid totals, and what scaled scoring actually means for test day.


Here is a piece of trivia that trips up almost everyone looking at the JLPT for the first time.

Every level, from N5 all the way up to N1, is scored out of 180 points. But the passing scores don't scale the way anyone expects:

  • N1: 100 / 180
  • N2: 90 / 180
  • N3: 95 / 180
  • N4: 90 / 180
  • N5: 80 / 180

Yes, you read that right. You need more points to pass N3 than you do to pass N2.

It's not a misprint. The pass mark is set separately for each level, and N2 happens to sit right at 50%. But clearing the total is rarely where people actually hit a wall.

The real problem is quieter.


The Sectional Minimum: How 95 Points Can Still Mean a Fail

You can score well above the passing line and still fail the exam.

Each test splits the 180 points into sections, and every section carries its own cutoff:

  • N1, N2, and N3 have three sections (60 points each): Language Knowledge (Vocab & Grammar), Reading, and Listening. The cutoff is 19 out of 60 for each.
  • N4 and N5 combine things into two sections: Language Knowledge + Reading (120 points, minimum 38) and Listening (60 points, minimum 19).

Miss a single section cutoff by one point, and you fail. No rounding up, no appeal.

Say someone takes N2 and gets 50 in Vocab/Grammar, 28 in Reading, and 17 in Listening. That adds up to 95 points, five above the 90-point pass mark.

Result? Failed. Listening came in two points under 19.

This hits self-taught learners hardest, the ones who study mainly from textbooks and flashcards. You can grind Anki decks and drill grammar patterns at your desk, and your Reading score will climb reliably. Listening doesn't build itself that way. Neglect audio practice, and that single 17 is what ends the attempt.

One extra mechanical detail worth noting: test periods do not equal scoring sections.

On test day, N3 is split into three timed blocks (Vocab; Grammar & Reading; Listening). But when your score report arrives, Grammar is bundled with Vocabulary into "Language Knowledge," while Reading stands on its own. If you track mock exam mistakes in a spreadsheet, sort them by scoring section, not by the test booklets.


Why You Can't Just "Count Correct Answers"

When taking practice papers, people naturally score them like a school test: "There were 50 questions, I got 35 right, so that's 70% of 60 = 42 points."

The real exam doesn't work like that.

The JLPT uses scaled scoring based on Item Response Theory (IRT). In plain English: questions don't carry fixed point values. Each question's difficulty is measured statistically, and scores are placed on a common scale, so an N2 pass earned in July 2026 means the same level of ability as one earned in December 2020.

The catch? The organizers don't publish their conversion tables or question weights.

So nobody, no prep book, tutor or forum post, can tell you your exact score from raw answers. The best you can do is an educated estimate.

The standard approach is proportional estimation: scale your share of correct answers to the section's maximum. And here's the rule of thumb I always give: leave yourself a safety buffer.

If your practice tests keep hovering around 20 or 21 in Listening, you are not safe. You are in the danger zone. On the real exam, nerves, fatigue in hour three, or a harder set of questions can easily drag a borderline 21 down to 18.

(Shameless plug: I got tired of doing the math by hand after every mock paper, so I built a free JLPT score calculator on my site, OK JLPT. You enter how many questions you got right per section, and it estimates your score and flags any section that sits too close to the cutoff.)


What Does a Pass Actually Mean? (A CEFR Reality Check)

People often ask how the N-levels compare with the European CEFR scale. The official JLPT site does not publish a score-to-CEFR conversion, so treat any table that gives exact score bands with suspicion. The rough, unofficial correspondence most people use is:

  • N5: around A1
  • N4: around A2
  • N3: between A2 and B1
  • N2: around B1 to B2
  • N1: around B2 to C1

A pass tells you that you cleared the pass mark and every section minimum, and nothing finer than that. Two people with passes at the same level can sit far apart in real ability.

Also, keep your expectations realistic: the JLPT only tests reading and listening. There is no speaking and no writing. Passing N2 or N1 proves you can read and follow Japanese accurately under time pressure. It doesn't automatically mean you can run a business meeting without stumbling.


After Test Day: Results and Validity

A few logistics to keep in the back of your head:

  1. Results schedule: July results go online in late August; December results in late January. Overseas results never open before Japan's.
  2. No rechecks: What you see is what you get. Scores are not rechecked once released.
  3. Validity: Officially, JLPT certificates never expire. Once you hold N2 or N1, it's yours for life. Some employers and schools set their own time limit, though, so check their requirements before you apply.

The pass marks, section minimums and result dates cited here come from the official JLPT site, jlpt.jp. The CEFR comparison is an unofficial rule of thumb. For quick answers to other test-day questions, I keep a source-cited JLPT FAQ on OK JLPT.

Top comments (0)