NEB10-beta usually works pretty well for this size. I am assuming you are looking to maximize the absolute total number of transformants you are getting and not actual "efficiency". Since efficiency is usually defined as CFU divided by transforming DNA in ug, optimal efficiency will be achieved in lower range of transforming DNA (for example 100pg-1ng for pUC19), however, the total number of output colonies can be increased by using up to 100ng of DNA but the efficiency will be lower. So bottom line is if you are looking for max number of colonies use neb-10-beta with 100ng of DNA or more (supercoiled form will transform best) and stick closely to the protocol, but if you are looking to actually calculate the transformation efficiency of your plasmid then use 100pg-1ng of plasmid as for a typical 50ul transformation reaction becomes saturated at 10ng of DNA thus using amounts higher than 1ng might make you efficiency calculation incorrect.
you can refer to this technical bulletin for more info: