Google shows a larger surface-code logical qubit outperforming a smaller one

Google Quantum AI/Willow and error-correction roadmap

Google measured logical-qubit performance across surface-code sizes and found its distance-5 logical qubit modestly outperformed an ensemble of distance-3 ones. It was the first sign that adding qubits for error correction could reduce errors rather than add them—the below-threshold breakthrough.

• Distance-5 logical qubit outperformed distance-3 ensemble—first evidence adding error-correction qubits reduces total errors.
• Below-threshold performance is the Holy Grail of quantum error correction, sought since Peter Shor's 1995 theory.
• Larger surface codes suppress logical-error rates exponentially as code distance increases.
• Foundation work for later Willow chip achieving full below-threshold scaling on 7x7 grid.