Anything you choose will be a compromise. At this size for some truly gut wrenching 3D stuff your optimum choice would be a flat plate airfoil. But if you need to be able to do the smoother and more open maneuvers then I'm thinking something more in the NACA0012 range or a Selig 8022 would be nice. But you're asking for two different performance tasks that have conflicting requirements. For good 3D you want to be able to stall the airfoil on demand to stop the airplane in mid air. For that the flat plate is king and anything else will be more stall resistant and thust harder to actually drag to a sudden halt.

Actually in my own flying with a flat plate aerobat the bigger more open maneuvers were actually pretty smooth. Where it'll all fall apart is during the higher speed portions where the extremely wide control surfaces needed to do truly great 3D stuff will begin to flutter. THAT is where the airfoiled wing with the greater thickness to give it more stiffness and strength comes in.

I know it seems low tech but the flat plate IS a valid option. And if your desire for 3D is serious then it's the best option in many respects.

Larger 3D models are getting away with "proper" airfoils but if you look at them the airfoils they are using these days often tend to be a thinner % thickness than a few years ago. And on a couple of models I've seen pictures off the airfoil's high point was noticably farther back than was normal a few years ago as well. Both of these traits help these larger models to act like they have a flat plate airfoil in the typical larger sizes.

As mentioned flutter is the enemy. If you truly want this design to have a really wide speed range that your dual mode requirement suggests then I'd recommend a Naca 0009 and sharpening the leading edge as required to achieve the "stall on demand" that you want for true 3D work. At the same time the typical 3D wide ailerons will be really wide and flutter prone. To combat that I suggest a slop free drive to somewhere near the middle of the control surface to mazimise the support and at the same time arrange for mass balances to bring the balance point of the control surface itself forward and close to the hinge line. Same with the tail surfaces. And all the movable surfaces need to be light but skinned with a layer of some sort that emphasises stiffness. Otherwise the segments of the control surfaces further from the control horn and mass balace will still flex and flutter on their own.