Diecast requires very costly tooling to make the dies. So to keep the price reasonable, you need a larger production run. With brass the work is done with simple fixtures and jigs, the tooling cost is relatively very low. Labor goes the other way, the brass is all manual labor while the big pieces on diecast are done as chunks.
Now with all the add-on details mentioned above, there is still quite a bit of manual work on diecast, but generally the major pieces such as boiler, with basic details, cab, tender shell, frame etc are cast and handed to the assembler. With Brass these would be fabricated and require more labor.
As mentioned the cost per unit for brass does not really go down as much for longer runs as does diecast. Brass when talking scale models vs toys however does still require the research and engineering to the get the dimensions right to make an accurate model. It seems with diecast, this is sometimes done, and sometimes not so much.
I don't know the numbers but I think with a brass steamer Scott can do a run of 100 or so while I think with diecast the mfg would need to talk in terms of at least 1000.
I asked one maker some years ago who was doing an HO PRR H21 hopper, if he would do one in O also. His reply was that he would start when I gave him a check for $75k. At that time that was he estimated the molds for injection plastic would cost. That was more than 10 years ago. So today it might be 125k-150k for a freight car mold. I think a diecast die/mold would be in the same price range. They both require detailed machine work on a block of steel to make a mold. So if a single unit is 125k a steamer and tender might be 250k. If we assume you could sell this for $1000, you would have to 250 just to pay for the tooling, before labor, design, packaging and all the rest - plus you have to make a profit. There are just big number involved.