Opossum is the first model-based optimization tool for Grasshopper and based on the RBFOpt library. It uses advanced machine learning techniques to find good solutions with a small number of function evaluations, i.e. simulations.
Opossum's GUI is similar to Galapagos. It's speciality is finding close to optimal solutions in 100 - 300 evaluations, while genetic algorithms and particle swarms typically need many more steps. (Various paper with benchmark results are availabe here.)
Opossum comes with a self-installing *.exe. To run it, you will also need a free license key, which you can get here.