I tried to open a module, but it is potted the hard way. It has a 1kW power dissipation, and I managed to break a module in that amplifier by running max. output current into a short. For sure, that module broke down due to too high junction temperature, but it took a while, so I don't think of internal thermal runaway but rather general overheating due to insufficient thermal sinking. The datasheet specifies some kind of SOA that covers the whole power dissipation at any combination of voltage / current as far as I remember. These old style modules are too old for the "many transistors" approach, I believe it's rather a single transistor die inside.
Regarding your amp: 600kHz BW from 150V rails is impressive, too
Its also possible that it handles linear operation better just due to being a BJT. Its the modern MOSFETS with low Rdson and IGBTs that are the most famous for developing hotspots in the die and blowing up under extended periods of linear operation.
And yes getting this much bandwidth from it has involved a lot of hair pulling. My first attempt was a BJT output stage design, but at the required currents they simply would not switch fast enough. Many many hours in LT Spice later resulted in a redesign of the whole thing and the use of a MOSFET output stage. It used low gate charge high voltage FETs and a strong gate drive circuitry to make sure it switches fast enough. All in all it ended up being made out of about 30 transistors. The gain stages, current sources, bias generation and gate drive was done with BJT transistors as they are fast enough at low currents. A few more transistors are used to switch the amplifier between standby and active within a few microseconds(Standby reduces power use to near nothing and also floats the output to eliminate the need for a Transmitt/Recieve switch later on)
The damn thing loved to oscillate in weird ways with all of that bandwidth. With certain loads at certain output power it would become a 400 MHz oscillator with significant output power even. I would never believe that a big power MOSFET can output a signal this fast. This particular oscillation turned out to be a LC circuit formed by the internal structure and tough hole pins of the MOSFET itself. Adding damping capacitors and ferrites in the right places calmed things down as well as putting ferrite rings on pins of MOSFETs. More LT Spice, cursing, some electrical shocks and exploded components later and eventually the thing was stable with all sorts of loads(Even heavily capacitive or heavily inductive and surviving dead shorts) and worked perfectly.
The resulting powerful output of the amplifier once ended up driving a pocketwatch sized lump of piezoelectric element so hard it exploded from the mechanical stresses.