That's going to depend upon a few factors. The applicable code. The general size range of materials. The welding process in use. The size of the finished weld and weld consumables in use. The specific joint design in use. Etc.
If working on very small parts with GTAW or Brazing then the flare bevel starts very soon with a very tiny radius.
BUT, if working with 3/16 7018 on a structural or heavy equipment application then the radius can be larger before it is considered a flare bevel groove weld instead of a fillet weld in a T-joint. If you want to get/be VERY technical, then when comparing ALL applicable clauses in A3.0 and A2.4 it would be a flare groove with the very beginning of any configuration that takes away the PURE 90° corner for a T-joint and makes any minute amount of bevel to the corner. But, there is normally a practical application to the pure technical application.
There are some applications that defy technical interpretation and are therefore handled by clauses in the applicable code that specify detail drawings of the exact joint and what the engineer expects for the finished product. Skewed joints are one of those we all deal with on a regular basis. Clause 2 in D1.1 explains the process in that case.
The question in my mind in order to guide one's decision as to if they are dealing with a flare bevel or a T-joint would be if the flare is large enough to make any significant difference in the weld penetration and finished dimensions for structural strength? But, what has the designer/engineer called out? Did they take the flare into account? Or just specify a specific fillet weld size? If they have called out a fillet weld and the flare is rather insignificant my personal opinion would be to just make the weld to the drawing dimensions and not cause a controversy. Probably not worth the added grief or the torment of arguing that your weld is large enough because of the radius and penetration.
Small radius members are often disregarded for engineered strength and only the fillet weld size has been taken into consideration. The inspector is thus only interested in the actual finish size of the legs of the weld and no consideration will be given by engineer or inspector for 'penetration' even with a flare bevel on the one member.
Edges of channel iron, angle iron, thinner wall HSS, and even the edges of some WF beams have a slight radius to them. But that radius is not going to be considered in the design or inspection of the part. Only heavier wall HSS with a pretty fair radius will be considered a flare bevel. Or, round bar, rebar, etc welded to a planar surface such that the OD of the bar is the radius and makes the joint a flare bevel groove.
Hope I made my explanations clear with some good examples to further explain.
He Is In Control, Have a Great Day, Brent