[...] several japanese brands (Samwa, Hioki) have excellent quality and reputation and also do not have markings. It boils down to the manufacturer's reputation and seriousness.
No- Sanwa, Hioki etc. will "design to" 61010 but never tested, no approvals, so it's another gamble.
Do you work at one of these companies? If not, you can't possibly know that.
Can we stop shilling multimeters from manufacturers such as Sanwa and Hioki, that have no 61010 safety approvals, only misleading statements such as "designed to" or "conforms to" or "safety rating" or "CE" which mean nothing.
Only Uni-Trend UT139E and S have 61010 certification, the UT139ABC ETL certificate page has a bad URL, surely a little mistake. I've asked for the formal certificate.
As far as the protection circuits with GDT/spark gap, I see most are across the inputs in series with the usual 1.1kohm PTC, like in the 34401a, PM300, and with the extra 1kohm surge resistor for the Keysight handheld DMM'S.
The antique Beckman/Wavetek/Meterman/Amprobe Tech300/3000/HD110 family used 2kV GDT or spark gap directly across the multimeter's input. A bunch of Radio Shack models as well- these are all from pre-1010 era.
Only Fluke had a spark gap after the HV input resistor, back in the day when they used that instead of MOV's.
So an ionized gas tube would give different fault currents, some explosive due to follow-through current, and others much PTC smoke when on a DC bus (where GDT normally extinguishes at zero-cross on mains) where it's a long-term wiener roast due to the tube lighting up and keeping current flowing in the PTC.