SAE standards for steel bolts 

The SAE has established a sequence of grades from 0 to 8 for steel bolts, on the basis of the metal from which the bolt is made and the manner of manufacture. Available grades run from 2 to 8, with 8 the strongest. Higher grade numbers almost always mean increased strength (an exception is that some grade 6 bolts are stronger than grade 7). The heads of steel bolts are marked to identify their grade.

It isn't always a good idea to replace a bolt with a stronger one.  Some bolts are deliberately chosen so that they are weak enough to fail before the stress or strain damages some more expensive or critical part of the equipment. For the same reason, in making furniture cabinetmakers use glues that are weaker than wood.  That way, if the furniture is overloaded, the joints break.  It is much easier to reglue a broken joint than to replace a piece of broken wood.

In the 1980s, large numbers of counterfeit bolts appeared in the United States, almost all imports. For this reason, the SAE grade markings can no longer be trusted unless one knows exactly who made and graded the bolt. Aerospace-grade bolts are also being counterfeited (even NASA has been duped, to the tune of one million dollars to disassemble the Astro 1 space lab to remove counterfeit and defective fasteners).

Torquing bolts

As a rule, when a bolt is installed, the nut (over a washer) should be turned and not the bolt's head. Unless a torque wrench is used the tendency is to undertighten large bolts and overtighten small ones. Suggested torques are given below. These suggestions do not apply if the bolt or nut has been specially lubricated.  In all cases, the specifications of the manufacturer of the equipment should be preferred to tables of approximations like these.

Torquing machine screws

Experiment is the best way of determining torque values for small fasteners. Pick at least 7 sets of parts at random and assemble them, torquing the fastener until it fails.  (The values at failure should fall within a fairly narrow range; if they don't, something is amiss with the parts, the torque wrench, or your technique.)  Find the average failure value, then take 65% of it. The following table gives a rough idea of values to be expected.