You said "If you are a deontologist, “the ends don’t justify the means” and it’s rarely ok to just kill someone to save someone else. If you are a consequentialist then the choice to live in the better world is obvious." and I took that to mean that killing someone to save five people was what you consider the consequentialist 'obvious choice'. From there you now clarify that the people are all the same so all we have to go on is quantity of living people, it seems. If "more people alive" is not what you mean by "the better world", then what do you mean by it?
You said “reducing it to more is better”, which is itself an oversimplification. That is not the principle at work here.
There are various utility functions that utilitarians might choose, for example hedonic (Mills) or eudaimonic. And various other assumptions you must make, such as “is the average life net-positive in utility or net-negative?”. Then you put your values into your utility function and assess the possible worlds and their likelihood of occurring. (For simplicity I’m leaving out the utility of having rules as heuristics, but worth mentioning as two-level utilitarianism seems appealing to me).
All else being equal, as in this thought experiment, saving more lives is better than saving fewer in my and most utilitarians’ opinion. But “more lives is better” is not the underlying principle at work in making the evaluation/decision. It’s merely the result of the calculation. One can easily construct thought experiments where more lives in existence would reduce utility and not increase it, in which case utilitarians would advocate for not saving those lives. For example if we modify the trolley problem to say pulling the lever will also cause the five survivors to be imprisoned and tortured for the rest of their lives.
(Look up “the repugnant conclusion” if you want to see a thought experiment where common intuitive morality breaks down under most systems).
