What do you call "the thing that produces results that surprise you because it wasn't what your model predicted"? I think of that thing as "reality"...not sure how you can coherently talk about a model without talking about the thing you have a model of?
Once you've accepted that reality = reality and that to discuss reality, humans always need to revert to models, then any further discussion of the nature of reality seems like a dead end. Because any way you try to discuss it is a simplification. Maybe it's more accurate to just accept that we're firmly in the realm of models and forget about trying to define reality itself. So in that way it becomes kind of pointless to discuss the nature of reality.
would "surprisal" work?
Only if you think all forms of realism are indefensible. But why think that given the success of science and technology?
That only describes the success of the scientific method in building ever more accurate models of the reality.