Richard is, of course, correct, but just to verbalize a little more... so, the "class" is not a rating, it is the characteristic of the output stage design. There are several defined classes, but A and AB are considered "linear"; that is, there is relatively little distortion of the input signal when amplified and that is why they are appropriate for audio amplification (AB also used for SSB radio, if you are interested). Generally that is because the initial operating point (determined by bias in tubes and base-emitter current in SS devices) of the devices is set so that all output current change driven by the input signal occurs in the linear portion of the device's characteristic curve (a property of each device type). Class A is the most linear and its initial operating point to accomplish this is such that there is higher no-signal (idling) current. Other linear classes do also have no-signal current, but less than class A. Some trade offs in the design choices for equivalent devices are that class A proves least distortion but lower power (because no-signal current is closer to max current) and probably hotter running because average current is therefore higher. I hope this doesn't conflict with any more authoritative source. :)