For GPUs, Only choice is X11, or Quark.
There is also Blakecoin (BLC) - which is the fastest coin (SHA3 finalist, Blake-256 algorithm).
3.4GH/s on a AMD Radeon R9 290X
The concern there is that it may be relatively easy to create professional mining hardware for. If GPUs move to the BLC solution they may not get to stay there very long. What we want is to bring in the GPU miners and keep them. Having 11 different types of hashing makes us much more secure in this way. We can argue that X11 gives long term safety.
The idea that ASIC resistance is important for crypto currencies has become widespread, however it is based on a fallacy. The premise is that ASICs are expensive and therefore make mining impractical for the average person. However, the production of ASICs is inevitable if a coin becomes popular enough, as is the case with Litecoin. Although Scrypt did a good job of stalling the creation of Litecoin ASICs, they are soon to be released. The properties that made Scrypt ASIC resistant are now the same properties that will make Scrypt ASICs exorbitantly expensive. Therefore in the end, trying to make a coin resistant to ASIC implementation only serves to make it more infeasible for the average person to mine it once ASICs are produced. This is why Blakecoin has been designed to be ASIC friendly. The simplistic nature of the Blake-256 algorithm makes it easier and cheaper to implement in silicon, so if or when Blake-256 ASICs are manufactured, they will be considerably less expensive and also considerably more power efficient than SHA-256 or Scrypt ASICs. GPU mining is excellent because it provides the flexibility to mine just about any coin, but ultimately it is a test bed to determine which coins are good enough to warrant production of dedicated hardware. One more advantage of Blakecoin in this regard is that due to it's support for merged mining, the efficiency of an ASIC would be compounded since it will be possible to mine multiple coins at once.