According to a 1990s study I found an earthquake around the level of the 1934 quake could be expected to have a death toll of 18,000-20,000. So this is what we could expect, although I worry it will be far higher. I don't know how many buildings there are in Kathmandu but they estimated 53k would be destroyed while reports are saying half have been.
Meanwhile an advocacy group cites a study that claims:
Sounds disturbingly close to what has happened.
They won't necessarily be similar in count. With magnitudes, there is a load of difference between even small variations in power. The amount of energy released in a 8.3 earthquake is 3.16 times that released in a 7.8. Even then, it also depends on where the epicenter was and how deep it occurred with respect to death counts.