The Julia General registry is locally stored as a tar.gz and has version info for all registered packages, so I tried this out for Julia packages. The top 5 are:

    DiffEqBase                  6.189.1   
    LoopVectorization           0.12.172  
    Reactant                    0.2.161   
    Mooncake                    0.4.159   
    Distributions               0.25.120

So, no crazy numbers or random unknown packages, all are major packages that have just had a lot of work and history to them. Out of the top 10, pretty much half were from the SciML ecosystem.

Caveats/constraints: Like the post, this ignores non-SemVer packages (which mostly used date-based versions) and also jll (binary wrapper) packages which just use their underlying C libraries' versions. Among jlls, the largest that isn't a date afaict is NEO_jll with 25.31.34666+0 as its version.

▲ dotancohen a day ago | parent | next [-]

You might want to try a different storing strategy. 0.25 is above 0.4. These are, I believe, what are called in Unix flags "human numbers".

▲

Savageman a day ago | parent [-]

I understood the list is ordered by biggest number, aka 189 > 172 > 161 > 159 > 120

	▲	Ghoelian a day ago \| parent [-]
		I think in semver 0.4 usually means 0.04, not 0.40..., so it should be lower than 0.25. Edit: nevermind, I misunderstood your point

▲ int_19h a day ago | parent | prev [-]

This would seem to imply that the vast majority of Julia packages are 0.x?

	▲	sundarurfriend 2 hours ago \| parent [-]
		There are many that are, but I feel like your comment is based on the same faulty assumption as your sibling comment - that this is an ordering of version numbers as a whole. It's not, the ordering is on the same basis as in the post, the largest single number within the MAJOR.Minor.patch trio.