For example you wouldn't say a plant has free will. It just sits where it is planted, collecting sunlight, water, CO2, and minerals, and grows as specified in its genome. It doesn't deviate from the evolutionary strategy of its species.
You probably wouldn't say an ant has free will. It can move around, sure, but it seemingly does so according to a complex set of rules that are consistently embodied by all members of its species.
Does a dog have free will? I think we're getting closer to the line. A dog can clearly make choices against the simplest evolutionary strategy. A dog needs food to eat, it's driven to prefer certain foods... but some dogs can choose to let a tasty treat sit on the table (or even on its nose!) without consuming it. Some dogs seem to act in a way that we would call altruistic, endangering itself to save a human child, for example.
And of course, humans are capable of making all sorts of decisions that seem counter to a simple conception of an evolutionary strategy. On the positive side, choosing to forego reproducing in favor of pursuing some social goal, like a career. On the negative side, choosing to kill someone despite knowing they could be caught and punished (and therefore not reproduce).
And I think the opposite holds too--when we imagine a being without free will, the defining characteristic seems to be that it always follows a rigid set of rules, with no possibility of deviation.
I keep referring to the "concept" of a simple evolutionary strategy; does that mean this is not a rigorous definition of free will? Well I would say that free will is itself a human concept, so I think it's valid to explore its meaning in the context of other human concepts.
I think it's entirely possible that humans are actually deterministic and predictable, but the rules of our behavior are so complex that we could never know or apply them in a practical way, the way we know and use the rules that govern plants, for instance.
Whereas myself... I know people in general can have a variety of responses to any given stimuli. But I know exactly what response I will have to a stimuli that has just occurred. From my point of view, I am very free-willed! After all, I'm "choosing" one out of a multitude of possible responses to this stimuli. But really, my "choice" is determined by physics, and the illusion of free will is exactly that I -- as the sole entity with the capability to observe my own internal state -- can predict my response, and no-one else can.
I think also, an additional factor is that of volition: the actions I take should somehow derive from, and act to fulfill, what I consider to be my (chemically deterministically determined) desires. Otherwise I might perceive "someone else" to be in control of my person.
But, if we were to speed your brain up a million times, when you look at my brain, it might seem a lot more predictable - much more like an inert machine than a dynamic animal. So there's a kind of relativism to whether one subject perceives another subject as "alive" or free-willed, given their own context.
Conversely, is the judgement that another entity is free-willed no more than an admission that one is unable to predict that other's actions, though that other can predict its own?