It doesn't require that much imagination to say that solar cells might one day be work in an extremely similar fashion to plants. Not likely, but not an outrageous thought.
Nature has produced a cheaper, more ubiquitous and more self-replicating solar system using efficiencies in the 5% range with a theoretical cap of 11%. That suggests we don't need 25% efficiency to accomplish amazing things. It isn't a critical metric.
The biofuel is burning what amounts to the plant's surplus energy that it wasn't using for anything, and recovering some of the energy that went growing the mass of the plant. It isn't comparable.
The point here is plants are covered with tiny green solar panels that are grossly inefficient compared to what humans produce. However, they are beyond cheap to produce (in fact they grow themselves) and suggest that we are not even close to pushing the limits on what we can do with solar energy design wise.
Efficiency of the solar panels really isn't all that important compared to making something with the flexibility and weight of a leaf. Comparing efficiency between solar panels is a waste of time outside the research community; all that matters is total cost to install vs. watts produced.