My hunch is that it might be possible to create something like https://polymathprojects.org/ for mathematical physics and I'd like to know whether MathOverflow users can recommend some appropriate problems. As Abdelmalek Abdessalam pointed out in the comments below such a platform can potentially be used to develop research programs in mathematical physics that are suitable for polymath type collaborative work.
The gist of this project would be to use collaborate problem-solving to tackle open problems in mathematical physics. Timothy Gowers wrote a seminal blog post on the subject here: Is massively collaborative mathematics possible?
There's a good point made by a commenter below as to why not simply use the current polymath platform. Well, I think that it would be easier for such a platform to succeed and gain support among mathematical physicists if it was maintained by mathematical physicists and focused mainly on problems in the area of mathematical physics. This is not the current setup of the Polymath project.