So, dragons, gryphons, what have you. They're usually (before Skyrim and Game of Thrones in the dragons' case) depicted as six-limbed, quadrupedal creatures. Now, the thing is, staying aloft is much easier than getting into the air. If I could cheese the takeoff, I could basically cheese flying.
According to a 2013 paper from Michael Habib, evolution often constrains the maximum size of flying creatures based on how they take off. Birds, for example, have a mass-inefficient launch scheme, as they use their legs to get the ground clearance for their wings. Those legs become dead weight during flight.
Bats and giant pterosaurs, however, can take off using their flight muscles and wings to pole-vault into the air. This works for about wingspans of 11 meters, the predicted wingspan of giant pterosaurs.
For heavier dragons, we could alter the wing's shape to have a higher surface area at the expense of aspect-ratio, which might not be a big problem.
The aspect ratio determines aerodynamic efficiency, as a high-aspect-ratio (read: very long and narrow) wing can generate more lift at lower speeds by separating a longer tube of air around it. However, due to how long it is, this wing-type requires lots of muscle to move because of leverage, and ridiculous ground clearance, forcing the bird to taxi, and get to sufficient speeds that way.
We are looking for a middle ground here, which comes on the wings of eagles, pelicans, and stonks.
Slotted, lower-aspect-ratio wings with a big area are preferred. While they require some speed, they don't suck at gliding, require less ground clearance, and the slots can
enslave harness the wingtip vortices.
But, here's the thing, assume we made the wings, they span about 11 meters and have a realistic wing loading (25> kg/m^2). Got all that? Well, we still have the issue that launch strategies are morphology-specific, and my dragons and gryphons have a lower profile than the quetzalcoatlus (i.e: they aren't as tall as a giraffe), plus an extra pair of lighter limbs, optimized for burst-speed.
So, now I'm here, thinking if there's a way to come up with a sensible launch strategy that can work with a creature that's less vertical than a giant pterosaur, and has six limbs. It's supposed to either exploit the creature's body plan, work in spite of, or regardless of it.
What should that launch strategy be?