Yes, 2Hut8 is right, but it's more than that. The thing that many overlook is that heat transfer works in three distinct modes; Convection, Conduction and Radiation. In the safe under fire, these modes are all active and contribute significant energy to the safe interior. The biggest, and most influential energy path to the interior is the metal jambs and door frame. These materials are heated to full external temperatures on the outside, and conducting heat thru the door opening perimeter structure. This all releases energy inside the safe by convection and IR emission. If the steam is not circulated inside the safe and venting thru the jambs, this energy drives the internal ambient temperatures up very rapidly. These complementary energy transfer effects are happening over the entire safe body to a significant degree, where corners and edges get very hot and conduct energy thru to the inner walls.
In addition to all of that, door seals are never perfect. When you heat a steel body to temperatures well above 1000ºF, there is substantial thermal expansion of the structure. This expansion is not consistent, nor controlled. There is no resilient seal material that can cope with the active changing shape of the structure. The body can bulge as much as 3 inches on the sides, same with the door face. Intumescent seals can expand quite a lot, but they are brittle and fragile once they expand, and they crumble easily with mechanical changes in the jamb shape and profile. Corners are particularly hard to seal, and the intumescent seals generally have joints in this area that don't close well. Door seals are trouble, and gaps are always present regardless of the seal approach. There are no commercially available flexible materials with adequate resilience to survive these high temperatures. The bottom line is that sealing the door is virtually impossible, and hence even a small leak of hot gasses from the outside will get in the safe and drive temperatures to failure levels very rapidly.
So, contrary to the notion that a steam saturated shell around a sealed inner container would be a good fire barrier, the complexity of ALL of thermal forces and structural issues in play make that thinking no more than a hope. I have tested safes with well sealed interior structures, and they never do well. We must exploit the cooling effect of the steam inside the safe, as well as the expanding nature of the steam gasses, to get anywhere near a controlled sub 350ºF interior environment. People have been fighting this science since the very first fire rated safes were conceived in the late 1800's. A lot of very smart people have tried to come up with a better mouse trap. Sorry to say, there are too many uneducated manufacturers that sell their untested theories to make claims that are destroyed when real fire exposure is imposed.