No, that's not how that works. You put in basic physics, which is known from experiments, and then there are assumptions about which effects have to be included

No, that's not how that works. You put in basic physics, which is known from experiments, and then there are assumptions about which effects have to be included - start with the big ones such as solar irradiation and ground IR radiation, and atmospheric stratification, and fluid dynamics and knowledge about air currents. Then add the ocean as a heat reservoir and ocean currents, ground albedo changes due to cover change (snow, desertification, crops...). Add some clouds in different heights. Oh, and stratospheric, tropospheric and ocean chemistry. Then add aerosols which influence cloud cover, vertical temperature distribution and thus winds, ... Of course, some variation in the solar irradiation due to changes in the sun and the Earth's orbit. And I probably forgot a lot of influences here which are already considered. Other important non-physical model parameters are things such as spatial grid size and time steps. Atm we are adding more influences to the models. It is thought that they have smaller impacts, but they might not have. All these above are things based in physics and the basic rules have withstood many experiments. But - how about methane fluxes into the atmosphere due to permafrost melting, or due to ocean heating? We are also not sure about many fluxes in nature because they are very hard to measure. Satellites help., but... How much methane do cows emit? There are halogenes coming out of the sea ice, why and how much is it, and what does it do to atmosperic chemistry? What processes take part in the soil and how much of what is emitted over the whole globe? Does it change with changing conditions, how much? There are many important questions in the details here, and you actually have to feed all of that into the models. This is where the assumptions are, because we do not know all of these things, and many more. And then you see what the result is. A small increase in overall temperature is not just that, it changes the vertical temperature profile, which changes circulation patterns (among other things). And then you have a different distribution of temperatures on the surface. That's what the models say, and that is what we can see in reality, thus the models are not that bad here. Of course, there is still some way to go until we can trust the more detailed results, I'd say.