Update to my own post: I think I am making newbie mistakes, and for the help of anyone who finds this thread in the future, it is the human-readable raxml info, not binary, that should be submitted.
Furthermore, this info file must have been run through Siavash's reformat-info.py script (bundled with sepp) which (perhaps among other changes) adds a line starting 'Base frequencies:' that lists (wait for it) base frequencies. I'd gotten an error earlier that told me 'base frequencies not found in raxml-info' or something like that, even though that line was present, which is why I thought qiime tools import wanted the binary version of the info. In reality, the error was very likely my mistake of copying the pre-reformat human-readable info file to 'raxml-info.txt'.
So now I'm on to a different error (that also might wind up being another stupid mistake):
An unexpected error has occurred:
Invalid character in sequence: b'U'.
Valid characters: ['A', 'Y', 'V', 'D', 'R', 'N', 'M', '.', 'B', 'K', 'W', 'H', 'T', 'G', 'S', 'C', '-']
Note: Use lowercase if your sequence contains lowercase characters not in the sequence's alphabet.
Checking my aligned fasta file now...