The Washington Post has an interesting article on climate models which features Gavin Schmidt making a robust defence of their usefulness:
Put in the conditions on Earth more than 20,000 years ago: they produce an Ice Age, NASA's Schmidt said. Put in the conditions from 1991, when a volcanic eruption filled the earth's atmosphere with a sun-shade of dust. The models produce cooling temperatures and shifts in wind patterns, Schmidt said, just like the real world did.
If the models are as flawed as critics say, Schmidt said, "You have to ask yourself, 'How come they work?' "
Now last time I heard, the models could get into an ice age but couldn't get out again, so I'm not sure whether Gavin is being entirely straight with us here. Perhaps the models have moved on though, although one could still wonder if they could move so quickly from not being able to get out of an ice age to being useful.