Table value or nameplate?
Both, for different jobs. Conductors size from the code table; overload sizes from the nameplate. This keeps them apart and says which did what.
Motor circuits: two numbers, two different jobs
Conductors size from the table. Overload sizes from the nameplate. They are deliberately different sources, and swapping them is the classic error.
The table FLC comes from NEC 430.248 (single-phase) or 430.250 (three-phase) in your adopted edition — find the row for your horsepower and voltage. The nameplate FLA is stamped on the motor itself. They will usually differ, and that is not a mistake in either of them.
From the code table
From the motor nameplate
Which number did which job
How this is calculated
conductor ampacity = table FLC × 125% (430.22, continuous duty)
overload = nameplate FLA × 125% or 115% (430.32)
The overload percentage is 125% where the motor's service factor is 1.15 or greater, or where it is marked with a temperature rise of 40°C; otherwise 115%. Some motors permit a higher setting where the lower one will not allow the motor to start — that is a separate provision and this page does not apply it for you.
What this does not do. Short-circuit and ground-fault protection — the breaker or fuse ahead of the motor — is sized from a percentage table in 430.52 that depends on the protective device type and the motor type, and that is a genuine code table this page does not contain. Nor does this handle multi-motor feeders, which have their own rule about taking the largest motor once at the higher percentage and the rest at their table values.
Why the code names a source for each calculation
The table value and the nameplate value describe different things and the code uses each where it belongs.
The table FLC is a standardised figure for a motor of that horsepower and voltage. It is deliberately generic. Conductors and short-circuit protection are sized from it precisely because it is generic: the wiring should not have to be redone when the motor is replaced with a different make of the same size, and the installation should be safe for any motor that could legitimately go in that spot.
The nameplate FLA is the actual current of the actual motor in front of you. Overload protection is sized from it because overload protection has one job — protecting this motor's windings from running hot — and a generic number cannot do that job. A protection setting based on a table value that happens to run high would let the real motor cook.
So the rule is not arbitrary and it is not a trick question. Size the wiring for the class of motor; size the overload for the individual.
The gap between them is normal
An efficient modern motor commonly draws less than its table figure. It is tempting to treat the smaller nameplate number as the more accurate one and size everything from it, which produces conductors and protection sized tighter than the code intends. The gap is not an error to be reconciled — it is the margin the standardised value is supposed to carry.
Starting current is a third thing entirely
Neither of these numbers describes what happens in the first second. Locked-rotor current can be six times full-load or more, which is why motor short-circuit protection is permitted to be set far above the conductor ampacity: it has to ignore the inrush and still clear a fault. That combination — a breaker much larger than the conductor would normally allow, plus a separate overload device sized close to the running current — looks wrong to anyone used to ordinary branch circuits, and it is correct. The two devices are doing two different jobs, which is the same theme as the two current values.