The M8T's frequency output options aren't really comparable to a full GPSDO like the thunderbolt - it has an onboard TCXO and some hardware to get various frequencies out of it, but I don't believe it has something like a PLL that is configured to the output frequency or even an analog drive frequency tuning adjustment for a VCO, it's more of a fixed oscillator for the board and then a digital divider which deals with the error in a way that shows you that noise you're seeing. I believe the M8T will acquire GPS quicker and at lower levels than the trimble, as it has a much more modern receiver, but without the OCXO and disciplining mechanism, it doesn't have nearly the frequency stability or holdover.
I've done a fair bit of testing of PPS output accuracy, and the M12+ is actually one of the better ones[EDIT] Was thinking of the Resolution SMT, the M12+ is reasonable but does show its age[/EDIT], if I remember right (my data's on a different PC, I need to get around to actually posting it), in terms of raw timing error on a given pulse, but all of the uBlox modules perform almost identically for PPS accuracy back several generations (I think I tested an LEA 5T or 6T, several 7 (non T) variants, the M8T, and another?), and while it does stay rock solid in the long term, I think the adjustment step for the output timing is something like 20ns, so while it's always centered around 1s, the noise pattern around it is digital looking and never extremely tight.
If you want really stable frequency output from a GPS based system, you really want that PPS or other frequency output disciplining a VCO of some description, the extra layer of smoothing goes a LONG way in terms of stability and it seems like GPS modules that don't advertise themselves as GPSDOs never include it, even if they're made for "timing".