I stand corrected about the issue regarding tube preamps driving long interconnects.
To be more specific about potential impedance mismatches between preamps and power amps, the general rule of thumb is that the input impedance of the amp must be at least ten times the worst output impedance of the preamp. A problem is that tube preamp manufacturers don't usually tell the whole truth about preamp output impedances. If they simply provide a number without stating the frequency response at which that impedance was measured, assume the impedance is much higher in the low bass frequencies - a stated number of "400 ohms" probably refers to the impedance when measured at 1 kHz., while there's an excellent chance that it's 4,000 ohms in the deep bass. Taking, then, an output impedance figure of 4,000 ohms, the input impedance of your amp would have to be at least 40,000 ohms (i.e., at least ten times the worst output impedance of the preamp) in order to avoid matching issues. Tube amps typically have input impedances of 100,000 ohms, so they usually don't pose a problem. Solid-state amps, however, can be 600 ohms or lower (examples include vintage Rowland and many studio amps). If you are running a Goldmund preamp with a 10 or 20 ohm output impedance (or Rowland or darTZeel preamp with 50 ohm output impedance), you won't have to worry about mismatches. If you are running a tube preamp, however, you can face an issue, particularly if your amp is solid-state.
Ralph (Atmasphere) asks a good question - why is your choice limited to Audio Research and Levinson? Do you live outside of the U.S. or Asia, where other high-end brands are frequently not distributed? Both the Levinson and the Audio Research are fully differential balanced - do you want an ARC Ref 3 or Levinson 32 because your Levinson amps are balanced? Used balanced tube preamps available for approximately the same price as a Ref 3, and that have equal or better performance than the Ref 3, include the BAT 52 and the Atmasphere reference preamp (the Einstein comes to mind, too, but it will be more expensive and there are very few on the used market). Many people also prefer the earlier Audio Research reference preamps, as it is much easier to try different brands of tubes (the Ref 3 uses the 6H30 tube, for which there are only a couple of brands).
If you are willing to consider balanced solid-state preamps other than a Levinson 32, and with sound and build quality that is as good or better, I would consider the Rowland Coherence II (it uses a battery power supply, like the darTZeel and ASR) or the best Pass preamp from a few years ago. I have a Coherence II.
If you do not require a fully differential balanced preamp, then there are at least a dozen single-ended tube preamps available used that are as good or better than the Ref 3. As for single-ended solid-state preamps, there are very few truly great ones. If you can live without a remote control and if you don't mind dual volume controls, the Lamm L2 is superb (it uses a tube in the power supply, but the circuit is fully solid-state). Naim's reference preamp is also superb, but like the Einstein tube preamp, there are very few available used.
Trance: I just saw your last post. Be very careful of equipment reviews, as many are unreliable, including reviews in the well known magazines. Equipment reviewers do not perform controlled blind reviews in laboratories - their reviews are subjective and depend upon the other components in their systems, the type of listening room that they have, the quality of the A/C power supplying their systems, the type of music that they listen to, whether they judge equipment on the basis of vinyl or digital, whether they still have good hearing, etc. Also remember that audio reviewers often review equipment that has been "loaned" to them by the manufacturer or distributor on a "long-term basis", and review a lot of equipment of manufacturers that advertise in their magazines. I'm not saying that all reviews are bullshit, but in addition to reading reviews, you need to get to know people who work on high-end equipment for a living, get to know musicians who also happen to be audiophiles, get to know manufacturers, dealers and distributors, get to know audiophiles who have very carefully assembled systems installed in good rooms who listen to recordings of unamplified acoustic instruments, etc. Finally, be very careful of what you read on Audiogon - like the idiot who was talking about bass loss in the event of long interconnects driven by tube preamps, many, if not most, of the people posting here do not know what they are talking about - asking a question in an uncontrolled Internet forum is like screaming out a question in a mall parking lot.
Good luck.
To be more specific about potential impedance mismatches between preamps and power amps, the general rule of thumb is that the input impedance of the amp must be at least ten times the worst output impedance of the preamp. A problem is that tube preamp manufacturers don't usually tell the whole truth about preamp output impedances. If they simply provide a number without stating the frequency response at which that impedance was measured, assume the impedance is much higher in the low bass frequencies - a stated number of "400 ohms" probably refers to the impedance when measured at 1 kHz., while there's an excellent chance that it's 4,000 ohms in the deep bass. Taking, then, an output impedance figure of 4,000 ohms, the input impedance of your amp would have to be at least 40,000 ohms (i.e., at least ten times the worst output impedance of the preamp) in order to avoid matching issues. Tube amps typically have input impedances of 100,000 ohms, so they usually don't pose a problem. Solid-state amps, however, can be 600 ohms or lower (examples include vintage Rowland and many studio amps). If you are running a Goldmund preamp with a 10 or 20 ohm output impedance (or Rowland or darTZeel preamp with 50 ohm output impedance), you won't have to worry about mismatches. If you are running a tube preamp, however, you can face an issue, particularly if your amp is solid-state.
Ralph (Atmasphere) asks a good question - why is your choice limited to Audio Research and Levinson? Do you live outside of the U.S. or Asia, where other high-end brands are frequently not distributed? Both the Levinson and the Audio Research are fully differential balanced - do you want an ARC Ref 3 or Levinson 32 because your Levinson amps are balanced? Used balanced tube preamps available for approximately the same price as a Ref 3, and that have equal or better performance than the Ref 3, include the BAT 52 and the Atmasphere reference preamp (the Einstein comes to mind, too, but it will be more expensive and there are very few on the used market). Many people also prefer the earlier Audio Research reference preamps, as it is much easier to try different brands of tubes (the Ref 3 uses the 6H30 tube, for which there are only a couple of brands).
If you are willing to consider balanced solid-state preamps other than a Levinson 32, and with sound and build quality that is as good or better, I would consider the Rowland Coherence II (it uses a battery power supply, like the darTZeel and ASR) or the best Pass preamp from a few years ago. I have a Coherence II.
If you do not require a fully differential balanced preamp, then there are at least a dozen single-ended tube preamps available used that are as good or better than the Ref 3. As for single-ended solid-state preamps, there are very few truly great ones. If you can live without a remote control and if you don't mind dual volume controls, the Lamm L2 is superb (it uses a tube in the power supply, but the circuit is fully solid-state). Naim's reference preamp is also superb, but like the Einstein tube preamp, there are very few available used.
Trance: I just saw your last post. Be very careful of equipment reviews, as many are unreliable, including reviews in the well known magazines. Equipment reviewers do not perform controlled blind reviews in laboratories - their reviews are subjective and depend upon the other components in their systems, the type of listening room that they have, the quality of the A/C power supplying their systems, the type of music that they listen to, whether they judge equipment on the basis of vinyl or digital, whether they still have good hearing, etc. Also remember that audio reviewers often review equipment that has been "loaned" to them by the manufacturer or distributor on a "long-term basis", and review a lot of equipment of manufacturers that advertise in their magazines. I'm not saying that all reviews are bullshit, but in addition to reading reviews, you need to get to know people who work on high-end equipment for a living, get to know musicians who also happen to be audiophiles, get to know manufacturers, dealers and distributors, get to know audiophiles who have very carefully assembled systems installed in good rooms who listen to recordings of unamplified acoustic instruments, etc. Finally, be very careful of what you read on Audiogon - like the idiot who was talking about bass loss in the event of long interconnects driven by tube preamps, many, if not most, of the people posting here do not know what they are talking about - asking a question in an uncontrolled Internet forum is like screaming out a question in a mall parking lot.
Good luck.