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

The prevailing view is that quantum phenomena can be harnessed to tackle certain problems beyond the reach of classical approaches. Quantifying this quantum–classical separation and demonstrating it on current quantum processors has remained elusive. Using a superconducting qubit processor, we show that quantum contextuality enables certain tasks to be performed with success probabilities beyond classical limits. We illustrate this capability with the magic square game, a foundational problem that demonstrates a clear separation from classical performance. We further quantify this separation through the Kochen–Specker–Bell inequality violation, yielding χ = 5.59±0.008 . This work provides a method for benchmarking quantum processors based on contextuality algorithms, with further studies on many-body contextuality and scaling presented in a companion talk (Part II)