Bigger does not always means better and ultimately everything depends upon the requirement of the application if for example a natural sounding voice is need then of course neural TTS are the choice otherwise simpler rule based TTS like espeak are the better choice.