I made two native speakers take this simple kanji knowledge quiz


I took it myself too. Apparently, I know around 750 characters. My mother, a native speaker who finished Japanese middle school but graduated from an American high school and still lives in the US, knows around 2500 characters. My aunt, a native speaker who has lived in Japan her entire life and finished nursing school, knows around 3300 characters.

See how you compare. Do at least 30 for a somewhat accurate estimate.


Yeah the last time I did it the first 10 gave me 3500 but after like 80 it was at 3200 which was almost the exact number of kanji i had in my srs decks.

I tried it again…and I got 4200 after 50. I think it was really confused about my level though, because after I hit the 4000 mark it gave me 嘘 and 駐 lol. I’ve been studying 1級 kanji recently so maybe thats throwing it off. Funny enough the hardest one I got though, 鹽, came in over the 6000 threshold but I only know it because…its written on my bag of salt in my kitchen.

Overall, the amount of kanji I “know” by most peoples definition is probably a lot closer to the 3200 i mentioned earlier.


That was fun! It thinks I know 2012 Kanji. Which seems about right.

Roughly 2800, which makes sense since I’m preparing for Kanken pre-1 but also haven’t studied seriously in a while.

I got some wrong that I’ve studied in the past.

After about 50 it gives me 1250, according to wkstats I should know about 1150, maybe a few more from reading, but probably not more than a few dozen, so seems about right. The font itself is really odd, I suspect it’s not prioritizing japanese glyphs, so for example I’ve seen lines that were instead split up into “drops”.

Also at one time I said I knew a kanji and it went and decreased the amount I probably know, what a ripoff

I imagine if you’re studying for kanken level 1, it may skew things a bit. You may not be learning characters by frequency like you do with the joyo kanji. So if you know a cluster of kanji in the 5000 frequency range, this algorithm in place may think you know more than you actually do. That’s why it provides an estimate range, and your bottom one is 3215.

1 Like

While I got 1954 - 2172 (era where I live in), I feel that it is inaccurate, or at least less. What do you mean by knowing the Kanji? One meaning? Most meanings related to usage / vocabularies? One reading? Most readings? What about misreading?

I feel that for more common ones, at least an On reading should be known, not just Kun-related vocabularies. (And of course, some meanings must be known.) Though, what decides common-ness? As inexperienced as most answerers shouldn’t decide it.

At least, typing and comparing answers should help to better the calculation and reduce bias. I made a feedback to the repo, regardless. I don’t really understand the maths behind it, though.

I also understand that at some point, it would hard for less advanced learners or non-natives, to make it accurate. (How comprehensive is JMdict, for example?)

I personally gave myself credit for knowing it if I could identify a word that included it, with the correct reading, and the correct meaning of that word.


My definition of “knowing” a kanji is knowing at least one word associated with the kanji. Of course, this doesn’t fly with more frequently used characters, but it works for the less common ones. Using this definition of “knowing,” I read in one of my kanji books that the average native Japanese adult “knows” over 3000 kanji.

This wasn’t meant to be a serious assessment anyway. Just make of it what you will.


If that has a reliable source, I can’t say anything, but counting only words and excluding names, id be a bit surprised if the average was over 3000. If we are talking adults who read a moderate amount and up, then thats not surprising, but I suspect some of my coworkers who don’t read at all would be dragging down the average.

1 Like

I don’t know how you could reasonably exclude people’s knowledge of names from the equation.

It was a book by Wolfgang Hadamitzky and Mark Spahn, who I think are fairly reputable, as they wrote a huge kanji dictionary. I believe the 3000 includes names too, I should have specified.

Thats outside the topic of discussion, though.

Yeah then that makes sense.

Thank you for sharing this resource.
I was skeptical upon reading your post about the capabilities of JiKen in gauging one’s kanji knowledge with minimal input…but, I must confess, I was amazed when it pegged me pretty accurately (just as advertised) after I was about 10 kanji deep into the quiz! Truly impressive. :+1:


Did about a 100 and it’s around 2200, which I guess is about right.

At 200 mark, I know 1708; but I can easily tell I need to know a lot more (and better).

Got bored at 70 and got 2009. According to wkstats I should know about 1924 so it’s pretty close. 10/10 will do it again.


Apparently I answered 501 lmao I got sucked in. It puts me at knowing about 1510


I had time so did over 200 and it was really accurate…if WK says I should know 1800 I got 1850 on this site. Neat tool. at around 20 answered it was also pretty accurate at 1900 or so.


I’m not sure how many I did (maybe 40-50?) and it finished saying I knew 1571. I’m not all that surprised (though still disappointed) since:

  1. I haven’t specifically studied kanji in about three years
  2. I’m really bad at recognizing kanji out of context (i.e. not in words or sentences)
  3. I don’t do production, so I had trouble confirming to myself that I knew a kanji even when I thought I might
  4. I was really strict, deciding that if I couldn’t think of a word with that kanji before checking the answer I’d vote “don’t know”, even if I knew words with it after seeing the examples. (e.g. When asked about 阻 I was like “well that’s one of the ton of kanji read そ, but I don’t remember which one”, but if I had been shown 阻止 I would have known immediately.)

I even got several basic kanji wrong that I’m sure I would have gotten right if they were shown as part of words. Oh well. I probably will have to study kanji again at some point.