Point charge

Both AP electrostatics courses define it in the same sentence. EK 10.1.A.1.iv in AP Physics 2 and EK 8.1.A.1.iv in C: E&M: a point charge is a model in which the physical size of a charged object or system is negligible in the context of the situation being analyzed.

Read the ending. In the context of the situation. Nothing is a point charge intrinsically; a metal sphere is one when it is far from everything and is not one when another sphere is nearly touching it. The model has to be justified per problem, and what it buys you is a single unambiguous $r$ to put in $\lvert \vec F_E \rvert = k \lvert q_1 q_2 \rvert / r^2$.

One case where it is exact, not approximate. EK 10.3.B.1.ii and EK 8.3.B.1.ii say the field outside an isolated sphere with a spherically symmetric charge distribution is the same as the field of a point charge of the same net charge sitting at the centre. So for a uniform sphere, measured from outside, the model is not a simplification at all, and $r$ runs from the centre rather than the surface.

Two conventions ride along. The exam conventions box on both sheets sets the electric potential to zero at an infinite distance from an isolated point charge. For adding several charges as signed scalars, the AP Physics 2 sheet prints $V = \frac{1}{4\pi\varepsilon_0}\sum_i \frac{q_i}{r_i}$, while the C: E&M sheet gives the integral form instead.

A test charge is a point charge with an extra condition attached, and the two words are not interchangeable.