On an untimed test where you are given directions on how to solve progressive matrices there is no reason to fail the question besides the test being wrong or not following directions. How was iq ever perceived as being anything besides effort. The only way might be if modern tests don't have directions in which case it's just knowledge of the directions.
Even under your assumption that everyone would solve correctly every fluid IQ test given enough time, the test outcome measured as completion time would still tell us that some people solve it faster and some slower, so it is a measure of processing speed. By analogy: everyone but the severely disabled can run or at least walk 100 meters. That doesn't mean everyone is going to achieve an olympic record in that sport.
What lies underneath that (processing) speed... i.e. how much is it effort and how much something else is another matter. Experiments that have tried to motivate people financially (albeit with pretty small sums) haven't been able to improve people's performance in such tests. On the other hand, practice/training does improve results to a certain degree.