It's not really a big leap in thermal resistance at all.
First, the Lux V emitter was 8C/W, not 11C/w (the star was 11C/W)
Second, slapping down 4 dies is going to improve thermal resistance even if you don't change a thing about each die. In fact, Seoul wouldn't have to change a thing to go from 7C/W for a single die emitter to 3C/W for a quad die emitter.
Just look at the Lux V vs. Lux I. 8C/W (4 dies) vs. 15C/W (1 die), all using the exact same dies and attach methods. By having 4 dies, you have 4x the area for each watt to be transferred to the heat sink. In theory, your thermal resistance should be 1/4 what it is with a single die solution, but it's not because the slug has some thermal resistance as well, and all 4 dies are sharing the same slug.
With the Lux V/Lux I comparison, you can do some math to solve for each part of the thermal resistance (junction->top of slug and slug thermal resistance)
Let X be the junction->top of slug resistance, and Y be the slug resistance
8C/W = X/4 + Y
15C/W = X + Y
X=9.33C/W (junction->top of slug)
Y=5.67C/W (slug resistance)
Likewise, we can apply the same method to the Seoul P4/P7:
3C/W = X/4 + Y
7C/W = X + Y
X=5.3C/W (junction->top of slug)
Y=1.7C/W (slug)
The thermal resistance of an individual die hasn't changed. It's just now you get to count 4 dies thermal path for the same power. It's pointless to merely look at the thermal resistance, since by design, you will be running these at 4X the power of a Seoul P4.
If you were to wire together 4 Seoul P4s and call them a single "unit", the thermal resistance of that device would actually be 1.75C/W (7C/W / 4). The thermal resistance is lower because they're not all sharing a single slug...The power is spread out over 4 slugs instead of being forced through a single slug's thermal path.
A low thermal resistance doesn't mean cooling will be easy (since it says nothing about the intended design power), nor that the junctions will remain cooler (for the same reason). If you run this thing at 1.4A (350mA/die), then that's 4.62W. The junctions will be 14C above the slug. Take a Seoul P4 (7C/W) at 350mA, it's junction is only 8C above the slug. From a per-die perspective, things are actually WORSE with a P7 vs. a P4, as the junction temperature will be higher when per-die current is the same.
If you run this thing at max (2.8A), then that's 10W of power, and the junctions will be 30C above the temp of the slug. At 700mA (same per-die current), a P4's junction is only 17C above the slug.
10W is a lot of power to dissipate in a small area (as those who have overdriven Lux Vs will be sure to know). Cooling these will not be easy, by any means. You will have to pay special attention to the thermal interface between the slug and heat sink to minimize additional thermal resistance, since you will have 5-10W going across it, rather than 1-3W that we're used to for normal power LEDs.
