Counting the public OBS instance as part of openSUSE is like counting every PPA as part of Ubuntu. (There are also third-party package sources for Nix, btw.)

The OBS claim in that (untouched for 6 years) wiki article is also unsourced, with no information on the methodology. Repology uses a bunch of scripts to munge package names which does things like translate between distros, combine packages that have been split into dev/lib/dbg outputs, etc. It's open-source and pretty straight forward.

Repology is also not about measuring the number of packages, but the number of projects packages. Number of packages isn't really a meaningful figure.

NixOS' numbers are high in part due to some massive autogenerated package collections for various programming language ecosystems, like everything in PyPi, Hackage, and some other repos like that.

FWIW, I love openSUSE and OBS especially. Imo openSUSE has some of the best tooling of any 'old school' distro with traditional binary package management, and it was long my favorite for that reason before I ever switched to NixOS. I also think package availability is not a good reason to avoid openSUSE, since OBS makes building and distributing packages for openSUSE so painless. (When I ran openSUSE, I just maintained my own repos for missing odds and ends on the public OBS instance.) I think this is especially true with respect to NixOS, where I think users should generally be prepared to dive into some packaging work in order to make the most of the system.

But I think Repology is basically right about the overall count of projects packaged here. NixOS is experiencing huge, exponential growth for the past few years. It's one of the largest and most active projects on all of GitHub. The Repology numbers for Nixpkgs are not a fluke. The numbers for openSUSE relative to Arch and Ubuntu also line up with my intuitions about package availability as someone who ran all three in some combination for several years, running basically the same software on all of them and privately packaging whatever was missing on any given distro for my own usage.

If you want to dive in and do some work to try to line up 1:1 comparisons I'm at least a little interested! Maybe we could start with some kind of minimal openSUSE server installation and see if any packages go uncounted by Repology and how they're counted in Nixpkgs.

PS: Many of the best reasons to love openSUSE also apply to NixOS. Here are some that come to mind:

  - community driven, not dominated by any singular, huge corporate stakeholder
  - clever use of package management tooling to ensure higher quality packages, e.g., ensuring that dependencies must be complete by running builds in disposable or sandboxed environments
  - strong emphasis on automation
  - very easy to interleave and maintain your own custom packages with the base system
  - free (gratis) public infrastructure for hosting and sharing binary package artifacts
  - rollbacks for package manager actions!

There are lots of important differences, too, but I think openSUSE developers might really like some things about NixOS. (Users, too, although NixOS lacks anything like openSUSE's comprehensive graphical configuration tools.)