I would look at the cost and decide based on that. Im going to start by saying I havent done any testing of this but based on my knowledge of thermodynamics (I have a degree in Engineering) I would say its not going to matter to much. While length does allow for more contact and thus more cooling, there is a limit to the effectiveness. There is an equation, which without getting to into detail, there are 3 variables which effect the heat transfer, the heat coefficient (how well the heat is transfered), the area of contact, and the difference in temp. This gives the transfer of heat per time. So difference in temp is set depending on water, so is the heat coefficient. The only variable which adding more length will change is the Area.
The problem with adding more length is the more length you have the less effective the length is. The first 5 feet is the most effective as the temp difference is the greatest, as you move through the chiller its gets less and less effective. So by the last 5 feet you arent getting much change. It gets a lot more complicated based on the method of heat transfer and time in contact, etc. But the based idea stays constant
Based on my experience the wort reaches an equilibrium temperature fairly quick in the chiller, slowing down the flow will help maximize the effectiveness of the length to make sure that the output of the water is as close to the temp of the wort as possible and adjusting as needed. There has been many discussion on this forum about the flow rate and length.
I would recommend just going with the 25', but thats just me and knowing that the extra 5' isnt adding much. If someone with more hands on experience has something different feel free to add to what I have said. Again I havent tested the difference but from my experience with my chiller thats what I would recommend,
Now to get the full bang for you buck on those 25' running the water through a pre-chiller submerged in ice water to lower the temp will give a greater cooling efficiency.