Ted99,
First, since your motor is not just a simple induction motor but is powered through an adjustable speed drive, the relationship that I suggested doesn't hold well. That relationship is only practical when the motor is powered by reasonably "clean" sine-wave voltage.
Second, using a 125 hp motor with a 100 hp makes sense. The "electrical noise" (harmonics, nasty spikes, etc.) impose a greater thermal load on the motor, and the greater heat dissipation capabilities of the motor are needed to avoid overheating.
Third, relying on metering the power input to the drive to estimate the work input to the process by the mixer is not a simplistically accurate method. The "real power" input to the drive includes resistance heating losses within the drive and the motor, and these losses become increasingly significant at reduced power levels. Most likely, bearing and seal losses will be reasonly trivial, but you will have to verify the actual situation for your application.
Still, since your accuracy needs are not overly burdensome, you may be able to use input power measurement along with some realistic efficiency allowances to estimate the work input to the process.
I would do some research into the nominal performance characteristics of the motor and the drive to determine some reasonable basis for making these allowances. Since the motor is probably operating well below its "normal" power rating, its actual efficiency is probably significantly diminished from its nominal, full-load efficiency.
In most cases, the real energy savings provided by adjustable speed drives comes from operating their load at reduced speed. The diminished efficiency of the combined drive and motor at part-load normally pales in comparison to the "work not done" by not driving the load at full speed.