SpaceX demonstrated supersonic retro-propulsion
One of the traditional bugaboos about designing a Mars lander has been the problem of supersonic retro-propulsion. This is a problem since any reasonable aeroshell has a terminal velocity of several Mach. So you need either supersonic retro-propulsion or heavy parachutes to brake to subsonic first (and then subsonic retro-propulsion). The original Mars Voyager proposals in the 1960s used supersonic retro-propulsion, but JPL couldn't get it to work. So, when a half-size Mars Voyager was resurrected as Viking, a supersonic parachute was added. All subsequent JPL landers/rovers have used this design. I recall Rob Manning giving a presentation to us JPL interns in 2006 saying that supersonic retro-propulsion would be the best thing for Mars EDL, but that noone knew how to get it working.That was until last month. On the first Falcon 9 v1.1 flight, SpaceX demonstrated supersonic retro-propulsion on the first stage (that was the part of the recovery that actually worked). If SpaceX really have cracked the code on this, then they will be able to land much easier on Mars than any proposals since the Mars Voyager days.
I'd heard that payloads about the size of the MSL are about as big as can be done with the existing types of EDL systems, as the parachute becomes too big to deploy reliably or something. (I remember something like that). So to go larger, new EDL Systems are needed. Mostly likely, replacing the parachutes with more retropropulsion for deceleration.
I think there's a couple ways possibly around the issue if firing an engine into the thin Martian supersonic slipstream if that's a problem.One is to have a big jettisonable aeroshield, with retro engines protected behind it. When the engines are lit, the shield kicks off and impacts the surface. That way the engines don't light into a supersonic slipstream. I would think once they are lit and pressurized, the thin Martian air wouldn't be able to to enger the nozzle and be a problem. The exhaust plume expelled out the bottom should overpower and basically split the slipstream.
Another is something like RedDragon where the nozzles are tucked back just out of the slipstream and the heatshield creates a bit of a splash efffect in the atmosphere (see picture below as well). As long as those nozzles are tucked in enough, they shouldn't have supersonic Martian air coming into them I don't think. I would guess this is the concept behind why RedDragon's superdraco's are expected to work for landing on Mars.If so, you can scale that up with larger engines. Perahps bigger methalox enignes in a scaled up geometry of Dragon. It's not as efficient to have engines positioned like that because of their outward angle. But it should hopefully be feasibel for landing purposes. A Mars ascent vehicle of some sort can have engines directly under it for better efficiency for ascent.
Horizontal lander. It's a much better configuration for use after landing and has many advantages during EDL as well.
The problem with capsules is that you'd need huge diameters. The MSL had already a 4.5m diameter heat shield.You could go full-propulsive but in that case the payload mass fraction is rather pathetic. In the paper above the baseline vehicle is 60t in low Mars orbit, of which 8.7t are payload delivered to the surface. Of course those 60t need to be transported from LEO to LMO, which brings total LEO departure mass to around 350t (ISP of 350s).