Your title said "5G9" but I'm going to assume your typing reflects what you really want to know; that is, how does the 5E3 compare to a 5E9 (or 5E9-A Tremolux).
I have a 5E3 copy, and I used to own a '55 5E9 Tremolux. Six of one, half-dozen of the other, excpet the Tremolux has trem and a bigger cabinet. They both sound essentially the same, and the name "Tremolux" comes from "Deluxe, with tremolo."
The changed phase inverter was all about having a method of injecting the trem signal with causing audible pumping at the speaker when nothing is being played. Leo Fender had it patented. But a paraphase offers roughly the same gain as a split-load inverter plus its pre-gain stage ahead of the inverter. Both systems use two triodes and arrive at basically the same end goal.
If we're strictly talking abut capability, the paraphase is capable of a larger output signal swing than a split-load inverter, if both use the same B+ voltage. That's because the paraphase needs a given triode to only supply one output signal with that supply voltage, while the split-load has to supply 2 output signals and still have some voltage left over across the triode.
Either inverter can be made to work if you need only a relatively small output signal to drive your output tubes, which is generally the case with 6V6's (especially lower voltage cathode bias circuits) and EL84's. Split-load inverters can also work supplying larger output signals if care is taken with the design, and if adequate supply voltage is available.