Comment by perihelions

- "The first Long March 5 is classed as a full success on Wikipedia, I couldn't find info there about a failure (the second one did blow up)."

The Wikipedia entry describes it as "suboptimal but workable initial orbit", which I interpret as a partial failure (coming from a military entity that's universally opaque about its failings). They're not inclined for language like "partial failure" that we get out of transparent countries—contrast that first Delta IV-H, which also reached a "workable" orbit—just not the intended one.

- "However it was later found that nothing at all was wrong with the vehicle, it was a failure in the ground software, and an identical vehicle successfully carried out its mission 7 months later."

Also true of the Ariane 5 explosion: that was a software bug (unhandled integer overflow) in the flight control unit. The important part isn't whether it's hardware or software, but whether they got it right or not, before launch.