420MHz - 450MHz stepsize
Actually, as of a few years back anyway, here in California the 70cm Ham band (the 420MHz to 450MHz band to which you are referring) coordinators decided to go from a standard channel separation of 25KHz to a standard of 20KHz in order to accommodate more repeaters. Now that DOES conflict with the commercial and government land-mobile bands above and below (406MHz-420MHz and 450MHz-512MHz) which continue to use 25KHz channels with 12.5KHz (1/2 25KHz) interstitial splinters. Increasingly, now with the move to narrower forms of narrowband FM the interstitials are becoming the standard (12.5KHz) which might mean, though I am not sure, that newer interstitials for low power use might be squeezed in at 1/2 12.5KHz or 6.25KHz increments. But this is for commercial, government and military equipment. Most commercially available ham gear is not built to the tolerances and specs of the equipment available for commercial land-mobile usage.
To use the narrower channels effectively you at the very least require tighter local oscillator drift specs as well as tighter receiver IF passband filter requirements. This is partly (I am assuming) why when the California 70cm band coordinators wanted to reduce the 25KHz channel spacing to add more channels they didn't go with the commercial land-mobile plan of halving the 25KHz channel to 12.5KHz; rather they compromised and went with a 20KHz channel plan which added more channels while not severely overtaxing the capabilities of most ham radio gear.
From a scanner user point of view it gets a little confusing. A lot of scanners are set to 12.5KHz steps in the whole UHF band from 406MHz to 512MHz by default which can obviously handle 25KHz as well as 12.5KHz steps. But this will run into problems with the California 20KHz channel plan as some repeaters will be off center so you might get a repeater at 448.060MHz which would have to be tuned in at 448.0625MHz by the scanner. Most scanners, like their ham radio receiver cousins, have fairly broad IF filters so even a 2.5KHz offset would not be too noticeable IF the center of the LO is tuned correctly and hasn't drifted too much. This can be confusing to the user, though.
Now since most US 70cm band repeaters are at 25KHz channels with densely populated areas like California using 20KHz channels then using 5KHz steps in the GRE unit for this band is actually a good choice if you're going to choose one step since 5KHz will evenly add up to 20KHz or 25KHz channel steps - no problem! Now, outside of that band it SHOULD go to either 6.25KHz or 12.5KHz defaults (and I think it does - I don't own a unit currently).
That said, I agree that the FlexStep thingy really SHOULD allow the user to override search step sizes at will and easily. Maybe it only works on user programmed search band limits as opposed to preprogrammed service-search type bands?!
I have always been of the opinion that sticking limited step sizes in wide band capable scanners according to some current national "plan" is a bad idea and results in undue software load. For one thing "plans" change with changes in channel space requirements and with technology and manufacturing economics improvements. Also, often published "plans" can be confusing and error-ridden especially to outside engineers who are not directly involved in the commercial/government/military land-mobile inner circle of equipment manufacturers (like Motorola, Harris, etc.) so folks such as GRE, especially since they are based in Japan and have to rely on open documents which may not have all the updates and addendums are at a severe disadvantage. This gets even worse when things are in a state of flux such as the current switch to "narrowband" narrowband FM (hooo...don't get me started on this one LOOOONG story here!). An example is the problems with the GRE bulit Radio Shack Pro-96 wherein the narrowband stepsize was correct for the commercial 150.8MHz to 162MHz portion of the VHF High band but incorrect for the upper portion of 162MHz to 174MHz (approximately). The upper "government" portion started out similar to the UHF stepsize of 25KHz so halved that to 12.5KHz while the commercial/public service lower portion was way back at 30KHz then halved to 15KHz and now halved to 7.5KHz! Hence the confusion - I think you see!
They really should just give up on setting limiting step sizes with hard band limits and just go for global programmable sizes with 2.5KHz increments and PLL's with a at least 5KHz resolution if not 2.5KHz. Also, make their LO's tight with at least a good stable VCTCXO with user PROGRAMMABLE tuning offsets to allow the user to partially compensate for drift over time (by just hooking the unit to a stable source or at least receiving a known good station). With current technology this should not be that hard to implement nor that terribly expensive. I am pretty sure that the 500 and the 96 both have at least a TCXO on board so making it a VCTCXO which can be tweaked by user firmware adjustments shouldn't be out of the realm of possibility. I know - room for user error - this would be certainly a special super expert setting with locks or some such but, man, I'm telling you - extremely useful in the field!!!! I've dealt with enough drifting oscillators to know!
My two cents...sorry I got a little long winded! Anyway, your 5KHz in the 70cm ham band should be fine, at least.
-Mike