Why does somebody always want a rule? OK, I'll play along...
As a rule of thumb, I think almost everyone is comfortable filling 3AA cylinders or 3AL cylinders to 2200 - 2400 psi, including your local gas supplier. The gas houses, it might be noted, aren't all that fussy about oxygen cleaning and servicing at the pressure they'll fill, though they dedicate the bottles. There are still a dozen or so people each year in the US that die because of fires related to the use of medical oxygen and a few more welders that get roasted - this at 15 psi.
Many scuba shops will fill 3AA or 3AL cylinder to 3000 psi. Hey, we're a wild bunch of iconoclasts but we also (supposedly) are cleaning, servicing and dedicating the bottles. There isn't much concern about smoking as an ignition source while using them, either.
In some limited aerospace applications, much higher pressure are achieved. I've got Haskel that is rated for 5000 psi of oxygen which is serious scary. The local Haskel rep told me a story about watching one of them cook off years ago at an aviation repair facility - at those pressures aluminum doesn't melt, it burns once you ignite it. That's where those NASA-type protocols and inspections come into play. For what it's worth, NASA defines any 100% O2 atmosphere above 5 psi as hazardous, the result of the Apollo fire.
So the rule, if there is one, is highly situational. The more work you are willing to do to prevent fires, the higher pressures you can (probably) live with. The weak point in any scuba system is usually the valve seat - either the tank or the 1st stage - which is usually made of Teflon (or similar) and which begins to degrade at temperatures below 300F and will decompose at temperatures below 600F. Remember, you don't have to have a conflagration to have a serious problem, a blown valve seat or a lungful of poisonous gas is bad enough.
Let me ask: what do you think is a reasonable psi limit for scuba - and why? Personally, based upon years of seeing tanks, valves, regulators and even compressor filters burn up in high O2/high psi situations, my rule is I'll pump 100% to 3000 psi so long as I'm confident that the bottle and valve are clean, otherwise it's a no go or a lower fill. That's it for me but YRMV.
This is what titanium looks like when it burns in 100% oxygen at 8 psi:
Oxygen fire in an operating room:
