Playing with a model of your system in BeamBoy (free download ... try it!), a 10 inch span of 0.8" diameter steel rods will deflect ~0.001" with a 25 lb load in the middle. Deflection scales linearly with load, so you can add up your gantry components to see what static deflection to expect. Cutting forces will probably be 20 lbs or less in all directions. Acceleration forces will scale with the mass of the gantry+router, count on accelerations of about 10-20% of gravity, again in all directions.
If you can make your supports grab the rail to prevent side deflection, I would think 10" spans are OK. This is the tough part though, since the side walls of the pylons are thin, as you have found out.
Note that tightening your pylons to the base (wood?) will cause the wood to crush, changing the position each time and hence the position of the rail itself. Adding a 1/4" thick, 3" wide Al strip under the pylons, running parallel to the rails would prevent this crushing effect (less concerns about the underside, but big washers would help). As you can see, it is getting complicated, but doable.
After thought : instead of having long bolts (read : stretchy and bendy) passing through the pylon and base, consider short bolts that secure the rod to the pylon through hanger-bolts or the like. This puts the sideways forces directly on a short, stubby bolt, instead of long bolts and short thin sections of pylon wall. You can always shim the pylons to get the height right.
Paul Rowntree
Vectric Gadgets, WarpDriver, StandingWave and Topo available at PaulRowntree.weebly.com