Quantum–Classical Separation in Bounded-Resource Tasks Arising from Measurement Contextuality, Part II

Following the foundational demonstration of quantum contextuality on a few qubits (Part I), this work examines the utility of many-body contextuality for scalable quantum-classical separation. To this effect, we implement the N-player GHZ game on a quantum processor, and win the game with higher probability than would be possible with the best classical strategy. The prepared GHZ state shows genuine entanglement up to 45 qubits, without readout error mitigation. The range of this entanglement extrapolates to more than 100 qubits with readout error mitigation. Separately, we solve a 2D hidden linear function problem, and exceed classical success rates. This separation in the hidden linear function problem is shown to be retained up to 105 qubits. Our work proposes novel ways to benchmark quantum processors using contextuality-based algorithms, showcasing their potential to quantify coherent errors.