I think this is a case where utilitarianism is misleading. Don't dash yourself against the problem of "what is the correct preference-elicitation and preference-aggregation algorithm for all living things." Just try to figure out what your own preferences say about animals dying vs. you eating meat.

In my own case, I am basically fine with killing chickens - suppose you gave me a button that paid me fifty cents and killed a chicken (somehow avoiding externalities). I would press this button all day. On the other hand, I would not press this button if it killed a pig - the payout would have to be much bigger ($50? $200?) before I'd do it.

Once I know about this preference of mine, I can let it affect my purchasing habits appropriately.