Bryston is great, and the Innersounds are also a good pick. Of all of these listed here in the thread, I have to 2nd the Pass amps and to say that as far as I can tell they are in a whole other league compared to most of what is mentioned here.
Innersound will have many more watts/$, but not quite the same sound in the mid and high frequencies. Bryston and Innersound are probably right on the inflection point on the curve of value/$ - the perfect place if you don't want to spend megabucks but want very serious gear.
That said, you can always pick up a used Pass Amp here or at RenoHiFi. However - be very careful. For Maggies, you only want to look at the .5 series (the most recent incarnation). So, for example 150.5, not 150. Or XA60.5, not XA60. (If you have 8-ohm speakers then it doesn't matter.) I would think that a used 150.5 might treat you really well and would be on the lower end for the Pass gear.
Some people say with maggies slightly lower quality and higher watts results in a more pleasing experience in the end. I have to say, I recently hooked up a kilowatt amp for 2 months (of much much lower quality overall than what I run normally) and it did bring out a different feeling than what I usually run and has me trying to figure out how to upgrade to higher watts.
In that vein, huge Kudos to Spectron as mentioned above. And - I haven't heard this but do wonder about it - as an interim value/$ play it would be interesting to do the highest wattage NuForce, Rowland, or even the super-high-wattage Rotel. I think that would be the 1572? 500 watts/channel at 4 ohms? And not very expensive compared to Pass or Spectron or even NuForce.
The drawback on some of the Class D stuff is the upper-mid and high range. So, if you got the 1572 for now, then in 3 years (or 3 months :)?) get something like a Pass Labs XA30.5 (Class A amp - extremely rare these days), and actively bi-amp, with the 500W on the bottom end, and the XA30.5 on the top end. That would be absolutely awesome! And a good value for what you are getting. I looked at the Rotel distortion curve vs freq several months ago, and the problems start well above the 200 - 250 crossover point that you would use on the 3.6s. The curve is totally flat below 250 Hz, can't remember, but I think it was 450 or 550 where it started moving up.