This came up in our user survey multiple times. The main concern is quality: Savchenko solutions are peer-reviewed by 72 contributors. Expanding to Irodov would need a similar contributor base. We could start with a separate section and see if the community forms around it.
It seems, all "Kvant" problems are solved except for new ones. I belive, this also applies to other mentioned problems collections. If we will consider fun element, there are "unsolved" Kapitza's problems in physics, they are "open problems" like #5.6.8 and #3.3.35 of Savchenko. "And if you don’t ever learn pieces of physics just for fun, then what’s the point of doing physics at all?" (Kevin Zhou). In addition, there may be A.R. Zilberman's problems books without solutions (Zilberman's problems are perfect).
I can fully support @emixter's opinion --- focus on solving the problems from Savchenko, which are unique. "Entia non sunt multiplicanda praeter necessitatem" (Occam's razor). Savchenko is a great textbook, excellent solutions for studying physics can be written. I have check out CPhO's problems on https://cn.pho.rs/ --- well, its seems like a sport or IPhO with maximum calculations; if its relevant, it will required a lot of physics athletes:)