The earths atmosphere adds to the cost of ground launch (drag) and subtracts from the cost of GEO->LEO transfer (aerobrake). Instead of 4km/s transfer vs 7km/s direct, it should be 2km/s vs 9km/s.
So my point still stands. The space elevator will only provide ready access to geostationary orbit. Journeys to more useful low orbits will benefit marginally at best.
There's a lower point on the cable where you're already in an elliptical orbit with the perigee at the altitude you want.
For a target orbit of 400km, that happens at about 23.8Mm, 57% of the way up the cable. At that point you just have to let go of the cable and you'll only need a circularizing burn when you reach 400km: Δv of ~2,137m/s.
This is not taking into account the further savings from aerobraking as I don't know how to calculate that.
The delta-v to transfer from GSO to an LEO at 400km is 4km/s, whilst to go from ground level is about 7km/s.
That's a significant fuel reduction, _and_ that fuel doesn't have to go up all at once: you can keep fuel tankers at the GSO station and fuel up a rocket from there.
https://en.wikipedia.org/wiki/Hohmann_transfer_orbit