You don’t need mlterm for Unicode; a regular xterm should suffice, provided your environment variables are set correctly.
I wrote a tutorial for this some time ago.
UTF‐8 support in the default ksh is weak. Typing and pasting characters work as expected, but both backspacing and cursor movement break horribly. mksh supports Unicode in emacs mode only (not vi mode… ☹). Bash should work normally.
I recommend setting up a compose file for Unicode characters that you use often, but this is impractical for other languages like Greek (beyond the occasional math character) or Japanese. Unfortunately, though I know the capabilities exist, I have never figured out how to enable them—X is woefully lacking in documentation. You may want to look into things like Anthy. I know that at least a couple OpenBSD developers are Japanese, so I presume somebody has it working.