Benchmark

Random Circuit Sampling Phase Transition (67 qubits)

Random Circuit Sampling (RCS) with cross-entropy benchmarking (XEB) · Other · 67 qubits · Cirq / Google Quantum AI stack

Benchmark extending the Random Circuit Sampling quantum-advantage demonstration to 67 qubits at 32 cycles. Google's team identified two XEB phase transitions: a dynamical transition with cycle count and a noise-induced quantum phase transition set by error per cycle. The experiment maps out the regime in which random circuits remain classically hard to spoof — even in the presence of realistic hardware noise — and provides an updated reference point for XEB-based quantum-advantage claims.

RCSXEBquantum-advantagerandom-circuitssuperconductingGoogle

Benchmark extending the Random Circuit Sampling quantum-advantage demonstration to 67 qubits at 32 cycles. Google's team identified two XEB phase transitions: a dynamical transition with cycle count and a noise-induced quantum phase transition set by error per cycle. The experiment maps out the regime in which random circuits remain classically hard to spoof — even in the presence of realistic hardware noise — and provides an updated reference point for XEB-based quantum-advantage claims.

Key Metrics
Qubits
67
Cycle depth
32
Benchmark metric
XEB fidelity
Phase transitions observed
2 (dynamical + noise-induced)
Why It Matters

Identifies the XEB phase structure separating classically-tractable from classically-hard random-circuit regimes, putting quantum-advantage claims on a firmer, falsifiable footing.

Hardware

Google Sycamore-class superconducting processor (67 qubits)

Framework

Cirq / Google Quantum AI stack