Don't attempt to calculate current, look to empirical
data. "Back in the day" we would bond post-to-post in
a package varying bond lengths and test with DC and
pulsed currents, to failure, and get fusing current vs
wire length (as in mermetic packaging, all heat has to
go out the wire ends, so long wires fail first).
The "maximum frequency" is not about the bond wire
alone, it has to comprehend the source and destination
circuits and the care-abouts that involve wire bond
inductance. A 50-ohm system would have one bond
wire inductance limit for a given loss@frequency, a
high-Z low-C input another, and the system says what
loss is acceptable.
"bond wire fusing current" would be a good keyword-set.