As I mentioned above..it doesn't need to get hot enough to melt stuff, just hot enough to soften the steel.
Yes and no. There is some truth to what you are saying. But since not all of the steel would be exposed to the extreme temperatures and the column would be of considerable mass my guess is a minimum of 10T per column. All that massive amount of steel would absorb quite a bit of heat before it was all hot enough to see any appreciable losses in strength, especially considering the margin of safety that would be built into the structure. So the way I see it you need to achieve really high local temperatures to possibly cause buckling in one of the flanges, or again, really high local temperatures so that it would be hot enough to be conducted through the column in that time period, rather than be disippated by the shear mass of the structure and all connected parts.
At work on the stay ring flanges where we weld steel anchors that are about 5 inches thick, we have about a two inch fillet to weld. It takes about an hour to heat the area to be welded with a ceramic heater to about 500 degrees F. It takes about 100 passes to weld. I've gone and felt the steel structure after a guy has been welding all day, and I find it remarkable how well that giant mass of steel is able to absorb all that energy the guy has been pouring into it all day. Pretty remarkable stuff. Mind you, It's 1 guy welding per 15T of steel.
I'd be pretty curious how much energy a big steel column would absorb from the open blaze. My guess is it would take quite a bit of jet fuel to get there. Pretty interesting stuff either way. And I still think my gut feeling is somewhat right.