You're asking a complicated question with a lot of possible solutions depending on what you plan on doing.
What you need to do is work backwards from whatever ABV you want, and whatever finishing gravities you want, your target apparent attenuation (mash temp and yeast and the like) and figure out the OGs you want. From there, you can determine how much fermentable sugar you need to provide overall out of your grains, and size your grain bill accordingly based on whatever your expected efficiency. You can do that via gravity points- multiply your desired volumes by your desired gravities (either in Plato as is or truncated SG, ie 1.068 would be 68, in either case the math works the same way, just be consistent), and combine them (6 gallons at 1.068 and 6 gallons at 1.045 would be 6*68+6*45=678). That'll give you a representation of total sugar content. Then divide by the combined volume, 12 gallons, and you get 56.5. Design your grain bill around a 12 gallon batch at 1.056-1.057.
And then from there, I'd gather THREE groups of runnings, all equivalently sized, with a combined volume large enough to split into whatever preboil volumes you need to get both of your 6 gallons. From there, measure the gravity of each gyle, and you can use the same math to determine the blends you need to hit your preboil gravities. When I do mash normal beers, I typically do my initial mash around 1.5 qts/lb (and usually partigyles as well), which gives first runnings in the neighborhood of 20°P (I work in Plato) if you mash properly. My subsequent second runnings are then in the neighborhood of 6-7° (going off memory here), and then 3rd runnings 2.5-3°P (again off of memory). I typically get 92-93% mash efficiency this way, but I control my process tightly. Either way, you may get higher than normal efficiency (I'm normally 88% efficiency). And then from there, the big beer will typically end up with maybe 80% of the first runnings, 20% of the 2nd runnings, and 50% of the third runnings, and the smaller beer 20%, 80%, and 50% respectively. Although that'll depend on the ABVs I'm going for. I'm typically aiming for relatively similar ABV ranges (the big beer perhaps a bit bigger, the smaller beer a bit smaller).
Point here is that you might find the best results having some of the first runnings in the small beer as well (as that'll have the best malt flavor), but this'll also elminate a lot of the guesswork and help you get a more predicable result.
Braukaiser has a partigyle/batch sparge worksheet that you may find very helpful.
http://www.braukaiser.com/wiki/index.php?title=Batch_Sparge_and_Party_Gyle_Simulator