STARFIGHTER COMMAND · Leaderboard · γ₁ = 14.134725141734693

FULL 25-TASK GRID — ALL SCORES

CLEARED PARTIAL 0% / PENDING

Data note: Human scores = 100% on all 25 tasks (ARC-AGI-3 Public Demo, empirical). Frontier model scores are from published ARC-AGI-3 Public Demo results. EOSE ft09 score is EMPIRICAL — scorecard e5de1521-eef7-4ddc-a21b-60a64d9..., 6/6 FOLDABLE, MIFSF fold engine. All other EOSE scores are SIMULATED DATA — pending.

TASK / LEVEL	HUMAN	EOSE	OPUS 4	GPT-5	GROK 4	GEMINI 3.1	STATUS
ft09 LEVEL 1 · geometry-limited	100%	100% ✓ EMPIRICAL	0%	0%	0%	4.76%	CLEARED
sp80 LEVEL 2 · multi-path	100%	— SIMULATED	4.76%	1.36%	0%	4.76%	ACTIVE
r11l LEVEL 3	100%	— SIMULATED	0%	0%	0%	4.76%	ACTIVE
tn36 LEVEL 4 · Opus-specific	100%	— SIMULATED	3.57%	0.56%	0%	0%	ACTIVE
lp85 LEVEL 5 · symmetric	100%	— SIMULATED	2.78%	2.78%	0%	0%	ACTIVE
s5i5 LEVEL 6	100%	— SIMULATED	0%	0%	0%	2.78%	ACTIVE
lf52 LEVEL 7	100%	— SIMULATED	0%	0%	0%	1.82%	ACTIVE
sc25 LEVEL 8	100%	— SIMULATED	0%	0.2%	0%	0%	ACTIVE
sb26 LEVEL 9	100%	— SIMULATED	0.19%	0%	0%	0.21%	ACTIVE
ar25 LEVEL 10 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
bp35 LEVEL 11 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
cd82 LEVEL 12 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
cn04 LEVEL 13 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
dc22 LEVEL 14 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
g50t LEVEL 15 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
ka59 LEVEL 16 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
ls20 LEVEL 17 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
m0r0 LEVEL 18 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
re86 LEVEL 19 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
sk48 LEVEL 20 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
su15 LEVEL 21 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
tr87 LEVEL 22 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
tu93 LEVEL 23 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
vc33 LEVEL 24 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP
wa30 LEVEL 25 · DEEP	100%	— SIMULATED	0%	0%	0%	0%	DEEP

SIMULATED DATA — EOSE ft09 score is empirical (scorecard e5de1521). All other EOSE scores pending clearance runs. Frontier model scores from published ARC-AGI-3 Public Demo results. Human = 100% all 25 tasks (empirical).