I don't think it's physically possible that an ebike is more efficient than a normal bike. You take two otherwise identical bikes, one is operated physically, directly from the energy source (legs). The other is operated by a motor that's less efficient than the muscles in the legs, run from a battery that loses energy (in the form of heat) as it discharges, through a controller that heats up as it runs, the batteries heated up again as they were charged, by a charger that lost heat as it worked, running from power that heated up the transmission lines back to the substation where heat was lost, through more heated transmission lines and back to the power plant where more losses occurred.

The efficiency measurement of the muscles gets an unfair but unavoidable advantage from the fact that most of the base maintenance level of inefficiency is going to be paid anyway, because the person is alive whether they're riding a bike or riding an ebike.

Add in the fact that the more you ride, the more efficient you become, because your body becomes more and more efficient at using fuel the more fit it becomes.

Measurements based on how much Big Macs cost per calorie versus gasoline are kind of bogus IMO, we're talking about efficiency of movement versus calorie input here.

A lot of the efficiency of a bike comes from the fact that you can coast a lot of the time, so it depends a lot on the riding style of the rider. The faster the rider goes, the less efficient they will be since they're fighting air resistance which increases (I think exponentially, as a square?) as speed increases. I'd guess peak efficiency is probably at around 12 to 15 MPH on a diamond frame bike, probably higher in more aerodynamic setups.