Short answer is that its better to parallel first, serialise second, so option 2.
In reality, both designs are going to cause you headaches. At 300A and so many connections/small cables, you're gonna be in a world of hurt with unbalanced cells and unnecessary losses.
Best practice is to strip all batteries down to individual cells, then build 12 x 1S4P blocks and make that into a 12S4P battery.
On a phone, so can't really give too many details, but if you care enough to rebuild, post back and I'll answer in detail.