I'm aware that using engine_force isn't supposed to be an accurate simulation, but what unit does it use? Judging by the 'suffix: kg' in the source code, I assume it's in kilogram-force meters, but I'm not sure.

Well, that depends on you. If you made all your models to a size of 1 unit = 1m and use 1 unit = 1m as a distance measure, just like recommended in the documentation, and the standard gravity is 9.8 then it is logical that you continue using metric or better SI units all the way. The 9.8 would then be [m*s⁻²], a torque you apply in [Nm], and a force in [N].

Just be consistent throughout your whole game.

Edit: Ah, you mean the vehicle_body. Idk if they follow their own recommendation in there. They suggest to try a value between 25 and 50 for a vehicle with a mass of 1000. So, that does not sound like Watts, which would be the unit one expects. So far for consistency :-/

Actually, 25hp (around 19,000W, which is the unit one would expect) is pretty close to my first car that had a mass of 750kg and barely made 100km/h with a strong tail wind :-) But hps have more than one definition. So, I think, you just have to try, or start with the hypothesis that the unit expected is somewhat close to kW ...

    Pixophir

    You're probably right, I imagine not everyone is going for high speeds or power, and so 25-50 (as a hp/kw or even kgf) value is probably all you'd really need (but it is far, far too low for 1,200kg). I think it's closest to being in kilowatts but kgf could work too. I'll do some more experimenting and find out what it is closest to.

      Often times real world physics and values for speed and acceleration, especially for vehicles, seem rather slow and uninspiring in a game or movie. Maybe you'd have to pimp it up a bit.