Here's the whole idea in one breath: a scientific model is a simplified stand-in for something real, built to do a job. A globe stands in for the Earth, a map stands in for a city, the particle model stands in for matter you can't see. None of them is the real thing — and that's not a flaw, it's the entire point.
Start with a Map
Pull up a map of your suburb. It is flat, it fits in your pocket, and it is wildly, deliberately wrong: the roads aren't really coloured lines, the houses aren't little squares, and the whole thing is shrunk thousands of times. Yet it's brilliant, because it kept the one thing you needed — how the streets connect — and dropped everything else: the traffic, the smells, the people, the actual size. A map that kept everything would just be the city again, and be no use at all. A model is useful because it leaves things out.
Models Are Everywhere in Science
Once you spot the move, you see it constantly. A globe keeps the shape and layout of the continents and drops the size, the weather and the people. A shell diagram of an atom keeps how many electrons there are and roughly how they're grouped, and drops the fact that they don't really sit in neat circles. The particle model keeps "matter is tiny moving bits" and drops what those bits actually look like. A food web keeps who-eats-whom and drops everything else about the animals. An equation like speed = distance ÷ time keeps the relationship between three numbers and drops the car, the road and the driver entirely.
Every one of those is a stand-in. Each was built so a person could understand, explain or predict something they couldn't otherwise get at — usually because the real thing is too big, too small, too slow, too fast or too messy to handle directly.
Flick through the explorer and toggle keep-versus-drop on each pair. Same move every time: hold on to the few things the job needs, let go of the rest. That's where every model in science comes from, and the rest of this concept is just learning to do it on purpose.