Theoretically, you can keep increasing the voltage until the motor sucumbs to material stress failures and literally flies apart. So there is not really a "max voltage" for the motor. Of course, at the same time you will be dumping increased current through the coils, which will be heating up. At some point the coils will either melt and turn into a pile of slag, or ignite and release all the magic smoke. Actually, the only thing in question is which happens first: explosion or immolation.
Rather than try to determine the maximum voltage that the motor can take (before self-destruction), you really want to characterize the motor at a specific voltage. Since you retrieved the motors from a toy truck, just determine the normal operating voltage provided by the truck power supply (most likely a set of batteries). You will need to determine two performance points - no load speed, and torque at stall. Run the motor at the operating voltage with no load on it and determine the speed (use a tachometer if you have one, or juct count really, really fast while watching the output shaft). Use a clamp meter around just one of the motor leads to determine the current draw at the same time. Record the speed point on a speed/torque plot, and the no load current on a speed/current plot.
Then do the same thing when the motor has a load applied to it that causes it to stall. One easy way to do this is to put a short arm (12" is ideal, or 1" if the motor is very small) on the motor output shaft, apply the current for a short period of time, and incrementally add small amounts of weight to the end of the arm until the motor can just hold the arm horizontally without moving. Note the total weight applied (including the weight of the arm itself!) and the distance from the output shaft, and convert to applied torque (in lb-ft, oz-in, N-m or any other convenient units like furlong-slugs). Record that point on your speed/torque plot, and connect the two points. Ditto for the current draw at stall.
You now have the basic speed/torque curve for the motor at one voltage, and a matching speed/current curve. When a different voltage is applied, these curves will move up or down on their axes, but the slopes of the curves will remain the same. With this data, you can determine the speed/torque combination at any point on the curve, and the curve for any voltage for the motor. You can also determine the total mechanical power for any point on the curve (MechPower[in watts] = Speed [rpm] / 60 * torque [in N-m] * 2*Pi). Or you can just plug the values in to the great
motor calculation spreadsheet by Dr. Joe in the white papers section, which will calculate the curves for you.
-dave