One of the interesting things about Christianity is that it's not using a probabilistic uncertainty framework at all. "Belief" in Christ is not just some confidence >30% in the godhood of Jesus- in Christianity, one simply believes or does not believe. This is part of why Lewis' logical dialectic has an appeal in that culture; it's basically Aristotelian, accepting propositions as True or False (this is also one of the reasons Thomas Aquinas is so revered for integrating the two in the first place, if I had to guess).

But, while this is the accepted inside-view way to approach the question of Jesus' divinity, it is probably a flawed way to interpret material experiences such as miracles. Note that even the Catholic church uses a formal system involving evidence and testimony, and thus in practice has a kind of rough 'confidence interval' for the truth of a miracle. Basically, you're stuck with the standard of actionable confidence.

Try to think of a precise answer to the question, just as an exercise. When miracles are the load-bearing component of your religious belief, it's going to come down to the degree of confidence that you need. If 40% is too high, what about 10%? 2%? For that matter, try to think about your current probability estimates in concrete terms as best you can, and notice when a given piece of information lowers that or raises it.