Leaving biology entirely aside, it depends how much resistance the blade faces.
If it faces no resistance at all, which is sometimes how monomolecular weapons are portrayed, then that's because (by whatever means) it's not interacting at all with the body, so how can it do any damage? This is more of a "neutrino blade" than a monomolecular blade, so it's not very realistic, but it illustrates an extreme end of the scale.
So, let's say the blade meets some resistance. This means it's moving stuff around. We're probably assuming that it's sharp enough to "cut through anything" -- the resistance won't bounce it back or snag it -- otherwise it's a rubbish weapon because "weighs nothing" and "bounces off" are the weapon characteristics of a feather. When it cuts, it takes adjacent molecules in the proteins it meets, and makes them not-adjacent any more.
Proteins won't survive that. Yes, in principle, if both sides of a molecule were held exactly in place the chemical bonds would reform, but both sides aren't held in place, (most of) the body is not so rigid at the cellular level.
We can take the loss of a few protein molecules, there's plenty more about. So at this point we need to bring some biology in: how not-adjacent, and will the cells survive it? I'm not entirely sure but I very much suspect that they will not. An ovum of course can be punctured with a needle and survive, but a point entry is not the same as being cloven across an entire plane. Supposing that in just one smallish region of the cell, the two "sides" move relative to one another by just the thickness of the cell wall, then clearly it can't just join back up again once the blade is out of the way. IIRC cell walls are often under tension, so if it doesn't join up that means it completely comes apart.
The further these little bits of cell move, the more resistance the blade faces, but "cuts through anything" means that almost all the work it does is going into separating micro-structures, very little of it is going into moving larger-scale components that can survive being rearranged. That is to say, it could fit into inter-cellular spaces, but it won't do that, because it's so sharp it will cleave through a cell, not push it out of the way. However much oomph the wielder puts into it, that work is translated directly into micro-damage on a plane through the target's body.
As such, I think to a close approximation we can say the blade kills every cell it touches. Furthermore, anything under tension (muscle fibres, ligaments, tendons, downward-hanging appendages such as arms, blood vessel walls, the diaphragm) is cloven for the period of time it takes the blade to pass. Assuming we're talking a monomolecular blade, not a monomolecular filament, this is a substantial time in molecular terms, the width of the blade (let us say 1cm) divided by the speed it's moving (a few m/s depending on the wielder's choice). So perhaps more than a millisecond. How far do the "sides" move apart in that time under their tension, never mind any work done by the blade in moving them? The structure cannot re-form provided it moves enough to bring it out of range of the molecular forces holding it together. Which is really not very far at all to move in a whole millisecond.
For stuff under compression, and supposing the cut is not perfectly perpendicular to the direction of compression, then there's a shearing force applied by the source of compression (basically, the weight of the target) during the time the blade passes. Similar result to a lesser extent: things slide down slopes, and they don't have to slide far to get everything out of alignment.
This blade with width would also need low friction on the sides of course, otherwise there's a lot of resistance slowing it down without cutting anything, and you end up with it embarrassingly stuck. Worst case scenario, you have to twist it and pull it back out the way it went in, like some kind of medieval peasant!
So, a blade with width can be as thin as you like, the body's internal forces will do the job of ripping everything apart. A filament perhaps needs a bit more thickness: enough to move everything out of van der Waal's range would be plenty because that means every structure it meets is definitely disrupted. Of course you can kill with a less thorough job than that, and the thickness of the filament is just a lower bound on the distance it separates the things it passes through: actually it'll move molecules further than that within a random range according to the kinetics of the collision.
Finally, "monomolecular" covers a range. In some sense a flawless diamond is a monomolecular bludgeon. It's plenty monomolecular, it just isn't sharp. Your monomolecular blade isn't necessarily the thickness of, say a molecule of polythene consisting of a carbon chain with hydrogen hanging off it. It might of comparable thickness to the cutting edge of an extremely sharp metal blade, and still "act sharper" if it's stronger and harder than steel, and if the whole blade is that thin, not just the cutting edge. This is more than big enough to totally separate any cells or other biological structures it encounters, beyond any hope of them chemically re-bonding even at van der Waals range, at any speed.
So a monomolecular blade is extremely destructive, but the work to do that destruction does still have to be put in by the wielder because it manifests as resistance to the blade's motion through the target. It might not be quite the simple matter of slicing someone's torso in half with a flick of the wrist, that we sometimes see in fiction. Depending on the characteristics of the blade you may still have to put your back, or at least your arm, into the swing. Absolute minimum, put in enough energy to overcome the chemical bonding energy of the damage you do, otherwise the blade comes to a halt in the target.