Well, the question is flawed because it assumes a binary distinction between "moral" and "immoral" acts. In utilitarianism, actions are more moral, or less moral, not completely moral or completely immoral in absolute terms. Morality is identified by utility, and falls on a continuum.
Just for illustration purposes, suppose you have five actions available to you: A, B, C, D, and E, which change the utility of the world by -10, -5, 0, +5, and +10 respectively. Is E the only moral action you can take according to utilitarianism? Is it immoral to do action D?
E is the most moral action you can take, D is less moral than E, C is less moral than D, B is less moral than C, A is less moral than B. If you want to draw the line somewhere and say, "actions beneath this line are immoral, and actions above it are moral," you may do so, but utilitarianism says nothing about this line or where to draw it.
You might think - "let's draw the line at 0. Actions that improve the world overall are moral, actions that harm it overall are immoral. So A and B are immoral, C is neutrally moral, and D and E are moral." This is tempting. But the problem is that you are measuring each action by a change in utility relative to a baseline level of utility. Where is the baseline? You can draw the baseline wherever you like.
The truth is that A is not -10, B is not -5, C is not 0, etc. A results in some amount of utility in the whole world - perhaps A results in 12349087578 total utility, B is 12349087583, C is 12349087588, D is 12349087593, E is 12349087598. If we draw the baseline at 12349087598, then relative to this baseline, A is -20, B is -15, C is -10, D is -5, E is 0. If we draw the baseline at 12349087578, then A is 0, B is 5, C is 10, D is 15, E is 20. So it is not so simple as saying that actions above 0 relative to a baseline are moral. We can draw the baseline wherever we like, but wherever we draw it, it will be arbitrary.