ARC Prize reported o3 at 87.5% on ARC-AGI-1 in 2024; ARC-AGI-2’s 2025 reset shows how public proof can age into procurement risk.
ARC Prize reported o3 at 87.5% on ARC-AGI-1 in 2024; ARC-AGI-2’s 2025 reset shows how public proof can age into procurement risk.