Welcome to Gaia! ::

I don't know the answer to it, but I know it will depend on the chemical makeup of the bottle and design. Weather the bottle has elements that strengthen the glass or the producer wanted to stay cheap and use conventional glass ingredients. I would say overall there is not one answer to this question; there may be a range of pressures but not one single atm pressure. Also are you talking about internal pressure inside the bottle or external pressure on the bottle. I assumed internal since you said "... take before exploding..."

The problem with this question is that it is too imprecise. Glass bottle of what dimensions? The thickness of the walls is a critical parameter. How is the pressure applied? If more or less uniform external pressure, the bottle will be in compressive stress, while internal pressure would put it in tension instead. The compressive strength of brittle materials like glass can be several times higher than tensile strength, so this is important. Additionally, what kind of glass is it and of what quality? For example, the tensile strength of glass varies by more than a factor of two. Since glass is not ductile, it is very responsive to internal flaws, which accumulate over time. If design strength is considered rather than the nominal strength, as would be proper for safety reasons, a further drop of about an order of magnitude becomes realistic.

The formula poweroutage gave, σ = PD/(4t), is the longitudinal stress in cylindrical pressure vessels for which the wall thickness w≪D. The hoop stress is double this amount. This might not be very accurate if your bottle has very thick walls, but considering the unknown parameters of the glass, the loss of accuracy here should be the least of your concerns.

Finally, while glass already has an unimpressive tensile strength, it is even less tolerant of shear. The easiest way to break glass is to put it under a transverse load, even tempered glass.

Not to mention, are you talking about total plastic failure, or rather just the beginings of dislocations in the materials? You would have to look at a stress strain curve for the object to really find the point of what level of force would be needed to cause failure. If you look past "Hoop Stress" and such and go even further, you can find failure points far before Sigma would show, but it takes alot more calculations I really dont feel like typing out.

and on a side not, the bottle was cracked just before it exploded, but there was shrapnel.