I think we'll have to start a war to know it.

Been seeing some interesting numbers in the DE regarding this. I know that I've killed one enemy unit(not sure what type though), and it was reported that 500 enemies were killed. However, for foreign wars, DE is reporting casualty rates with numbers like 4,325 and 2,022 and 167 and 2,417. I'm suspecting that the unit model for this game is a bit more sophisticated than 1 = 1,000.

And I wouldn't be surprised if different types of units have different numbers of troops in them. One standard infantry unit could have more troops than a fighter squadron, for example. But that's just rampant speculation.

The number of soldiers in unit depends on its "Manpower" costs.
So 1 Infantry still equals roughly 1000 men.
This may change, though. I think 1 Infantry unit should rather be a division, not just a regiment.

infantry is 1500 men 1300 men in motorized infantry 1200 men in mechanized infantry

I watch my casualty counts fairly carefully, guys, and my comments here are based on those observations. If anything I say here is contrary to the rule book, I stand ready to be corrected by game administrators or more experienced players . . . .

My impression, based on my observations, is that the casualty counts take into account the battle readiness or combat strength ("morale") of the unit lost. If your conventional infantry unit is only 50% strength before the battle begins, and the unit is lost, the casualty count will only show 750 individual casualties (50% of a full-strength regiment of 1,500 men). Casualty counts also reflect units of varying sizes: conventional infantry regiments have 1,500 men, motorized infantry has 1,300; armored car brigades have 675 men, light tank brigades have 625, and medium tank brigades have 800; fighter squadrons have 400 men, strategic bomber squadrons only have 325; and so on.

Thus, the casualty counts are pretty sophisticated, as they account for unit size, pre-battle strength, and partial losses at the conclusion of the battle. One other observation: the unit casualty counts reported in the newspaper some times omit individual lost units for reasons I can't explain, but the casualty count by individual men appears to me to be pretty accurate.