One with a cosine corrected sensor and with an immersion factor applied will work just fine. You can get them with cosine correction, just the sensor has to be square to the source or else you can get bad output. It is easy to apply a immersion factor, if you need one, just do not forget to do it.
PAR meter captures the visible range of light from about 400nm to 700nm. Most LED get read pretty well by the recent iteration of PAR meters. They oversample some of the spectrum just a bit, but since nearly all their output is in the visible range, a PAR meter grabs it pretty well. PAR meters will under sample T5s a bit and MH a lot since they have useful spectrum outside of the visible range that a PAR meter cannot capture.
I have an Apogee MQ-510 as well... no complaints. It is cosine corrected, has the larger spectrum sensor and has the immersion factor applied already.