Only
Who will it be?
is correct. The issue arises from a difference between ordinary English grammar and prescribed English grammar concerning the case of complements of the verb "to be".
The verb "to be" is intransitive: it does not take a direct object. Instead, it can take a noun phrase as a "complement".
the "traditional grammar" rule is that the complement should be in the nominative if the subject of the verb phrase is in the nominative. In other words, traditional grammar advice prescribes "Will it be he?" and forbids "Will it be him?" See this Grammarphobia post: How should you answer the phone? Note that the traditional rule does not call for the complement to be in the nominative in all circumstances. The complement has to be accusative in sentences such as "I knew it to be him" where the subject of "to be" is an accusative pronoun like "it". It's somewhat unclear what case traditionalists would prescribe for the complement of "being" or "to be" when there is no obvious subject of the verb, or when the subject is in the genitive case.
native English speakers tend to put the complement in the accusative in all circumstances. That's why sentences like "Will it be him?" are commonly produced and judged as grammatical by native speakers.
In "Who will it be", "who" is the complement of "(will) be", not the subject. (You can tell because "it" is occupying the subject slot right after the auxiliary verb, and the sentence can't have two subjects.) So according to traditional grammar rules, only "who" is correct.
The rule calling for "whom" to be used in accusative contexts is not part of native English speakers' grammar, so the fact that we naturally use accusative pronouns for complements of "to be" is irrelevant. Modern native speakers naturally use "who" in both nominative and accusative contexts. This results in "Who will it be?" existing alongside "Will it be him?"
"Whom will it be?" is a hypercorrection: it isn't justified either by the grammatical rules that come naturally to English speakers, or by the set of rules that are traditionally prescribed. To get it, you have to mix-and-match, resulting in a sentence that sounds bad and that nearly all educated speakers would agree is ungrammatical.