The real improvement only arrives by pushing the point of lamina flow further up the RPM/flow range and delaying the onset of turbulent flow. For a given perfect tube you will reach turbulent flow at some flow level and throughput will fall drastically despite raising pump end pressure, lowering bearing end pressure, and causing more bypass (where the spring comes in). Until that things are fine.
Smoothing things up definitely helps push that point higher, too. But the existing tube is fairly smooth, and 1mm larger ID than the bored out gallery. The worst parts aren't the 13 to 11 to 13 steps, they're the sharp right angle corners in a few places.
And you stated you were interested in low RPM flow. Although it's true that any improvement will help push pressure drop down, final pressure up, and flow up with it, measuring the improvement at the lower RPM non-turbulent end of the flow spectrum might be a challenge
If you did bore out the lower gallery to 13mm and retap the end for a bigger bung then yeah, there'd be a benefit in overall flow (only significant if the stock system was breaking into turbulent flow at some point, though), provided you followed the improvements through the filter input and output (easy enough to do). The upper rail is bi-directional so wouldn't need much touching and can almost certainly out-flow even a 13mm pump-to-upper-rail connection:
PI * 5.5 * 5.5 * 2 = 190.1 sq mm
PI * 6.5 * 6.5 = 132.7 sq mm
Not taking the non-uniform flow rate, nor the non-even split in flow to the two halves, into account, that's a clear win to the upper rail over any lower feed improvements.
Max effort: Bore that lower gallery out to 13mm and retap the end! Then radius all 90 bends. Then rest assured you've got zero pressure drop issues at higher RPM