My preferences in order:
#1 but only on two conditions: (A) the problems encountered by others have been fixed, and the miners will operate stably at 1.2 Gh/s (and not kits, we are talking assembled ready to hash miners), and (B) they can be shipped in the next 7 days. Given that 1 Th/s miners with 1000 watt power draw are shipping for 2600-3000, this is a pretty big step down so I think you still need to consider compensation (see below).
#6 If #1 isn't feasible, and on the same conditions (miners are actually shipping in 7 days, and are stable 70 Mh/s). Again, I think comp is still an issue.
#4 if #1 and #2 cannot be done within a short period of time, and there really is no other good alternative.
As far as compensation, if AMT wants to continue to stay in business the above do not address compensation at all. Obviously, if this is essentially a liquidation plan then compensation is moot as you just be delivering what you can. However, if AMT is going to sell new miners (scrypt or otherwise), you owe it to those of us who (funded) stuck with you all this time to give back some form of compensation even if not just cash. Even if BTC price had stayed above 700-800, the increased in difficulty (predictable) and the decrease in cost per Gh/s (also predictable) has made the delay cost us more than 60% (e.g, A1-based 1 Th/s miners going for around $3K now plus the loss of mining March - May). Assuming AMT plans to stay in business, the compensation that would seem feasible could be: (A) additional or more powerful hashing boards provided at cost, when available, and (B) discounts on future miners (e.g., 50% up to $3,500 per 1.2 Th/s miner paid for).
We have a plan for that, but it might not be in the realm that you may be thinking. But lets just everyone that bought from AMT did get a form of extra compensation in the end, and you'd all be happy. First we have to get everyone mining, and once everyone is we're going to make sure everyone mines right.