I understand you want what you rely on to be mature, and battle-tested, but presumably they could use an algorithm used in OpenPGP like ECC or RSA and then re-encrypt that ciphertext with Kyber or something.
From my understanding this is what Signal and Tuta do, combine classic encryption for battle-tested protection but then also add post quantum encryption to defend against harvest now decrypt later.
OpenPGP is an IETF standard and that stuff is never bleeding edge. The workgroups don’t have robust hierarchies and they’re not in a rush. Discussions that take place over emails have high latency. All of it adds friction.
Anyway, the ball has been moving since Dec 2022: RFC 9980 - Post-Quantum Cryptography in OpenPGP
They’re now at the second-to-last maturity level. Once the proposed standard becomes internet standard, it might hit the next stone down hill and get it moving. There’s bound to be some PGP clients that are faster than others. But this is one gigantic generational step for the protocol so it will take time to get everything right and for it to start working for everyone will take as long as it takes the slowest client.
Signal is centralized small project that’s not nearly as top-heavy, and they don’t have to coordinate with a bunch of clients who in some contexts might have a say in how decisions are made, or at least right to be heard. Thus it’s a project that enjoys completely different level of security agility.
Tuta is probably running two separate protocols: in-house post-quantum protocol for Tuta-users and PGP for the rest. So there’s some agility there too.
Unless something has changed recently, I believe Tuta just uses their in-house protocol for everything.
I would definitely look into using age instead of PGP though. I’m not exactly sure what your use case is though but you could have a look through this What To Use Instead of PGP - Dhole Moments
I would also like to bring attention to this discussion: Replace GnuPG with Sequoia-PGP (& Actively warn against GnuPG)