Related thread: https://forums.libretro.com/t/regarding-the-choice-of-prboom/658
Imho, GZDoom would be awesome to have, since it has become a very capable engine able to support a lot of crazy extensions with many custom WADs out there that make use of them to make games that look and feel like its very own thing, and not just a doom mod. Also, it supports all the other Doom-based games, like Hexen and Strife.
Latest GZDoom got rid of FMOD so it's now 100% GPL-friendly in its license.
But I know it's a very complex engine and so it might be hard work to port. Something like Chocolate Doom seems like a simpler and very portable option. It's still actively maintained and because of its very specific goals it's likely to continue alive for a very long time. It has already spawn its own family of forks, like you mentioned.
However, there won't be much advantage in adding a Chocolate Doom core if the PrBoom core works fine. And from the comments in the thread I linked, it looks like the main advantage vs GZDoom (vanilla demo compatibility) would be lost when made into a core anyway.
So I'd still go for GZDoom personally, or maybe the Eternity Engine, which might be a good compromise (some of the features from GZDoom, while being also vanilla demo compatible).
I'd throw some money to the bounty of a GZDoom core if there's one.