Add extra task :

Can you find answer in O(1) time/memory complexity?

Using sum of Sum of an arithmetic series(progreesion)?

There should be a fixed test --> 1 across all languages as some solutions will have trailing lines as output (1\n)

C Translation (author inactive).

As stated in prior comment, the description suggests that only natural numbers will be passed in as arguments. Test cases show otherwise, which is confusing. It would be best to establish input expectations in the description.

Description should be language-agnostic

Invalid inputs are not being tested in the main tests.

m is excluded from the multiples

That's not tested and some wrong solutions pass by chance. And in some other languages, correct solutions won't pass.

Returning different data types is bad practice, consider returning Null or something similar for invalid values

The haskell fixture doesn't test if m is a multiple of n.

According to the description, it should be excluded, my solution (which does not exlude m) passed.

C# description says up to rather than including. Very misleading

As has already been mentioned, the written instructions should be updated so it's clearer as to what is expected or what kind of tests it will be run against.
It is not stated that the variables may be negative, or n may be > m, and need to be checked for validity (E.g. n,m > 0 && n < m).

The description is misleading: positivity is not an assumption, it must me validated for by your code.