I am developing a simple mobile puzzle game and I am currently working on the localization aspect of it. I want to localize it in Simplified Chinese, Japanese and Korean. I am planning to use the noto-cjk fonts collection, but the issue is the font size are very big due to the amount of glyphs used. Since my game doesn't have much text, I doubt I need all of those glyphs.
I have a way of creating a font subset of only the characters I use in my game, but I would like to have more than the bare minimum, hence the title of this question.
Where can I find the 3000-5000 most commonly used characters for each of those languages specifically?
Unicode has almost 94,000 CJK ideographs. But based on expert input from the China, Japan and Korean standards bodies has defined a subset called "II Core" with 9,810 ideographs that are considered (as of 2001) a minimal set required for East Asian markets. See http://www.unicode.org/reports/tr38/#kIICore for additional info.
There is also an updated subset that was defined in 2020, UnihanCore2020. See http://www.unicode.org/reports/tr38/#kUnihanCore2020 for additional info.
You can find PDFs with these repertoires at http://www.unicode.org/charts/unihan.html. This information can also be extracted programmatically from Unihan data files (in Unihan.zip) that are part of the Unicode Character Database—see https://www.unicode.org/ucd/.
These may be more than you're looking for, but you'll likely want some subset of the II Core set.