Like many of my other input island solutions, I've fallen back on the inefficient but simpler method of having blue do literally all the molecule movement and have red just handle bond timing and other support. But with three different bonding patterns, it was red that I barely fit in. This is another one I didn't at first realize was doable as an input island - I forgot that when you have enough bonders to put more than two by the output, it becomes possible to output maximally-bonded molecules after debonding them. I definitely wouldn't want to try this with anything less than 8 bonders though, theoretically possible or not.