Very seldom, if ever, true when you're talking about the BJT
as a saturating (i.e. any kind of efficient) switch. Raw small
signal bandwidth is another matter entirely.
BJTs have (generally) much smaller Cjc per Ic than a MOSFET
has Cdg for Id owing to the construction and physics of the
conduction. Just look at the conduction of a vertical bipolar
vs the MOSFET. BJT conducts orthogonal to the junction
plane and uses it all. MOSFET conducts parallel to the channel
plane and is velocity limited, cross-section constrained with
a whole lot of overhead plate capacitance that contributes
nothing to conduction.