the reason the htx-100 has a the narrower i.f. filter (3khz) is because it has no a.m. the 2510 has a 5khz filter because it has a.m. if you pass a 6khz wide a.m. signal through a 3khz filter, it is going to sound like crap. uniden could have used separate ssb and a.m. chains with filters in each chain (3khz for ssb and 5-6khz for a.m.), but then they would have to pass this cost on to the consumer.
this is exactly what cobra/uniden did with the 148gtl (used individual fitlers for a.m. and ssb). and to go a step further, they used a 6 pole crystal filter in the ssb chain which narrowed the passband down to about 2.1-2.4khz.
even the newer uniden clones (rci, galaxy, etc) use a separate filter for ssb that is about 2.5-3khz wide. this is about the same bandwidth as for the htx-100.
a radio being multi-banded has nothing to do with its selectivity. it is mainly in the design of the i.f. circuits as demonstrated by the htx100 and 2510 comparison.