There are currently three gender options: Gender-unknown, gender-female, and gender-male. This is a case of an English bias. It reflects the usage of the English localization, which really offers three different gendered forms. Many other localizations offer only two different gendered forms, or none. In those localizations, at least one of the three options does not have any effect to the users (while in that localization). This is bad UI design.
An ideal solution to this issue would be a localization-specific setting where translators would set the number of gender options so it corresponds to the respective localization's requirements. In accordance to this setting, users would always be offered the right number of choices of the currently active localization (or at least of the current wiki’s default localization).
Localizations that offer only two different gendered forms include the French, German, Russian or Spanish localizations. In these localizations, users should only be offered two gender options instead of three. Localizations that do not offer any gendered forms at all include the Chinese or Japanese localizations. In these localizations, users should be offered no gender options at all instead of three.
On top of offering superfluous options that do not have any effect, the current situation may have yet another adverse effect: If the messages are blindly translated from English instead of reflecting the respective localization, users may be mislead into believing that they have options that do not exist. Until recently, the German localization of the gender-unknown message promised users that they could choose not to disclose their gender and instead be referred to in a gender-neutral form, when in fact they were simply referred to in the masculine form. The French and Spanish localizations of the gender-unknown message promise users that they would be referred to in a gender-neutral form 'whenever possible', when in fact they are simply referred to in the masculine form. Of course, this can be improved by making the respective translations reflect the actual usage. But even then, users are still offered superfluous options that do not have any effect.
This issue is related to, but different from T61643. This issue is about reducing the number of gender options so it better reflects the actual current usage of the respective localizations, whereas that other issue is about increasing the number of gender options so some localizations might potentially be expanded in the future.