Is there an improvement braking from the hoods from Shimano Sora 3500 to Sora R3000?

The R3000 levers have the shifting cables routed under the bar tape as opposed to coming out the sides on the older 3500 levers. If you care for the esthetics then that might be another reason to upgrade. I'm not sure if the braking performance actually changed as I don't have older Sora to compare to, but the R3000 levers in my experience seem to offer good braking performance from the hoods. The design of the R3000 levers is similar to the other higher tier models (tiagra, 105, etc) so I assume that the braking performance would be the same as those, assuming the brakes at the other end of the cables were also of the same quality.

So my question is does anyone know if there is a significant improvement braking from the hoods between the Sora 3500 levers and the Sora R3000?

What would you consider an "improvement"?

The mechanical advantage of dual pivot brakes is a compromise.

Dual pivot brakes, with their forced centering, cannot track a truly wobbly wheel. If a spoke breaks e.g. on a 36-spoke wheel, the ability to track the minor wobble is dependent on the mechanical advantage. High mechanical advantage (which means you get lots of braking power with little finger force) would mean the wheel jams once per rotation. Low mechanical advantage could track a wheel with even a major wobble, but then again you need lots of finger force to brake.

If the mechanical advantage is high, brakes lose their "bite" with worn brake pads really quickly and you have to be continuously adjusting the barrel adjuster. Low mechanical advantage would allow you to never touch the barrel adjuster, but then again low mechanical advantage would require unacceptably high finger force.

So as you can see, either low and high mechanical advantage has their problems. Thus, there is an optimum between low and high mechanical advantage. With this optimum, finger force required is acceptable, the frequency of needing to turn barrel adjuster is acceptable, the ability of the brakes to track a wobbly wheel with broken spoke is acceptable.

If you go lower than the optimal mechanical advantage, it is worse than the optimum due to unacceptable finger force. If you go higher than the optimal mechanical advantage, it is worse than the optimum due to unacceptable frequency of needing to turn the barrel adjuster and inability to operate on a wheel with a broken spoke.

The optimum has been discovered decades ago. The optimum is time invariant. What was optimal 10 years ago is still optimal because there has been no "breakthrough invention" in ye olde rim brakes (whereas disc brakes have improved all the time).

About the last invention in rim brake technology was the move away from the faulty single pivot sidepull centering spring that couldn't center brakes reliably. In mountain bikes, the move away was via cantilever brakes to V brakes with their reliable centering springs. In road bikes, the move away was to dual-pivot sidepulls with their forced centering mechanism.

Most likely Sora 3500 and Sora R3000 systems both use the optimal mechanical advantage. Thus, no change. If you are interested in some example mechanical advantages, here is a resource I have written here.

If you have hands that are on the smaller side it can be difficult to get a good squeeze on the levers. Shimano does offer a shim that can be installed. The shim essentially makes the lever a few millimeter closer to the bar. When installing them you will likely have readjust the brakes as the pads will be closer to the rim.