One could change the gain to a higher value to make it work with an 100 mV input level. However, this will reduce the feedback and would incease distortion. The inverting type of circuit looks good on paper, but it need the stage before the amplifier to deliver some current and this stage will add to the distortion, especially at higher gain. So the proper way would be to add a stage before this amplifier to bring the 100 mV level to some 1-3 V level first.
The output stages with 3 transistors in a NPN/PNP/NPN (and complementary) combination with local feedback can be tricky. It can depend on the speed of the 3 transistors to have a suitable combination. So it is not only wattage to look at, but also speed. Finding suitable sets of transistors is where you might need the simulation for. Depending on the transistors used it might take some extra pF range caps to stabilize this. This is likely the really tricky part with this type of amplifier. The nasty part is that also parasitic properties of the layout may come into play here. The shown circuit is one of those that can give pretty good performance if done right, but could as well end up as a MHz range power oscillator if a different brand of caps / transistors are used.