As long as that box was chosen at random, and not pre-selected to be one

without the prize. If the latter you have a variant of the Monte-Hall problem,

and the probability that the box you have contains the prize is still 1/5.

RonL

(A Bayesian analysis also gives the p=1/4 result confirming intuition,

but note 1 in 5 times the box opened will contain the prize)