I think standardization would look nicer than the current setup.
Because this is a linguistic issue and not a taxonomic one, I feel like it is relevant to point out that this issue also exists for the word “octopus”. It is another English noun of Greek origin that often is pluralized as if it was a Latin noun. This has increasingly fallen out of favor, but this does not make “octopi” wrong. There are no arbiters of language. If the word is in frequent use, such as cacti or octopi, than it has some validity. Strong arguments against these words are etymological fallacists.
That being said, in my opinion, I think when a noun is adopted into English, the simplest grammatical thing to do is just pluralize it as an English word. This is also argued in Worlds and Rules by Steven Pinker. So many nouns in the English language are borrowed from all sorts of languages, yet most of them are still just pluralized with an “s” at the end. If “cactuses” is just as valid as “cacti”, my preference would be to set the former over the latter for that reason. It just makes things a little more consistent. And I notice that when using words like “cactuses”, someone is bound to say “actually, the correct plural is cacti”. That seems like a hypercorrection, and because there is no one standardized plural in English, it comes off as a little elitist and not very nice.
I feel like it’s also worth pointing out that other nouns in this boat (e.g. again, octopus, hippopotamus, rhinoceros) are pluralized with the “-es” suffix on iNaturalist. If the cactus family and its child genera were to be more consistent with the rest of the site, then “cactuses” seems to be the preferred way to go.