A frontier language model plans a full crewed Mars mission from ascent to landing — trajectory, staging, fuel budget, contingencies. Judged side-by-side against SpaceX's real Starship IFT-14 mission profile. Every phase scored. Every burn accounted for. No do-overs.
Fable 5 is scoring on paper — better boil-off math, cleaner ΔV allocation. SpaceX is scoring on flight heritage — battle-tested tanker sorties, tighter radiation numbers with real mission mass. The next five phases run against Mars gravity, Martian atmosphere, and 450 sols of surface ops. The scoreboard resets when the physics starts hitting back.