I've approached this calculation is several ways in the past, but I think I found the simplest approach, via using the calorie estimator here: Calorie Calculator - MapMyRide

Using inputs for an average male 200lbs 5'8" 35 years. Riding 1 mile in 10 minutes:

71 calories burned, per mile.

Of course food calories are actually kcal's, and 1 kcal = 1.16222 Wh

71*1000*1.16222 = 82.5178 Wh/mi

So this is already at 1/3rd what a Model 3 achieves (~230 Wh/mi), a very large number!

Now, let's take into account thermodynamic conversion losses:

For the Model 3 powered with solar/wind, I roughly estimate "well-to-wheel" efficiency as 70%

For a human powered with food, I roughly estimate the "well-to-wheel" efficiency as 15%. Now obviously this is a difficult estimate, but any research into this area I've found has come up with a similar number.

So applying efficiencies:

Model 3: 230/.70 = 328.57 Wh/mi (1 passenger)

Cyclist: 82.5178/.15 = 550 Wh/mi

Cycling seems significantly worse just from a thermodynamics perspective, especially considering the Model 3 can take up to 5 people, which brings the per person consumption to 65.7 Wh/mi. Almost any way you slice that, 5 passengers in Model 3 beats 5 cyclists, but it doesn't even seem necessary to have passengers.

Of course, I didn't estimate production costs for the Model 3, but I also don't have a good way to estimate the food production cost, not just thermodynamically but holistically for the cyclist. There's water and arable land usage, as well as methane production, both at the animal as well as the human. Also, people that commute to work on a bike will often take a shower after their morning (please, don't stop) and evening commute, which increases their water consumption even more.