@RJT the best way to test any loco is running light, i.e., not pulling a load. The weight of the trailing consist, and in some cases its momentum, will mask potential jerkiness. For a real challenge, try running the loco from a straight section into a right circular curve* of minimum rated diameter. MTH says these new Pacifics should operate on O42 curves. But they were derived from MTH's very first scale offering way back in 1994, at which time they were rated for O31 curves. MTH wanted to sell trains, and O31 is what most folks had at the time, including me.
The problem is, these are heavy locos. Historically, MTH has gauged its wheels wide with little "wiggle room," and placed rubber tires on both sides of the rear axle. Thus, the wheels cannot skid to acommodate the difference in length between the inside and outside rails. (This is a much bigger problem on O31 or O42 than it is on O72!) Check your ammeter... When the loco enters a tight sectional curve, there is a rapid, significant increase in friction. The speed control has to really be on its toes or you will see a hesitation, followed by a surge. Watch closely. Frankly I would be surprised if any of your locos could complete this difficult test smoothly and realistically at 3 scale mph.
In the days before speed control, I experimented a lot, and concluded that the Pittman motor had to be turning at least 800 rpm to avoid a noticeable slowdown. USRA Pacifics, at least the first ones, were geared at 16:1, which is surprisingly toylike for what was billed as a scale model. The drivetrain was self-locking, and the "flywheel" was too tiny to confer any meaningful coasting. Eight-hundred motor rpm translates to about 10 scale mph. I was disappointed, because my uncle's postwar 773 could negotiate the O31 turns more smoothly, and at a slower speed. So much for progress! Slap on the band-aid of speed control circa 2001, and now you can get around that curve at 3-4 mph, but with a subtle hesitation and surge as described above. The preferable answer would have been to double the gear ratio, adding a bigger flywheel in the bargain.
Things might improve a little with break-in running, but given my experience, I wouldn't expect better performance than what you're seeing.
*We mostly use track systems where each piece has constant radius and makes up part of a circle. In real life, track curves have a spiral easement. If you have the space and are willing to mess with flex track, "eased" curves would create a gradual increase in friction and give the speed control system more time to react.