Try adjusting the malt bill so that the OG is the same across the board, then reading the IBU estimates from there. If they are still drastically different, there could be other factors coming into play with varying parameters of the software. The true test is to run the same recipe through the different software and comparing the predicted data of those to your specific measured results (OG, FG, SRM, efficiency, etc.). As long as you measured and entered your volumes and constants as accurately as possible, the software that comes the closest might be the one to work with from there on out.
Brewing software tends to be more accurate with better quality inputs. Even then you may have to determine which program suits your process better than the others.