A redesigned error-correction circuit on IBM’s 156-qubit Heron r2 processor raised logical qubit survival from below 90% to more than 96% per correction cycle.
Mid-circuit measurements emerged as the main culprit: while some qubits are measured during computation, others must idle, and that waiting time injects noise that undermines quantum logic operations.
The team quantified that measurement noise is now one of the dominant reliability limits on present-day superconducting quantum devices, giving engineers a clearer target for hardware and protocol improvements.
Nature Communications published the study, which stems from a 2024 University of Sydney-IBM collaboration funded by IARPA and aimed at advancing fault-tolerant quantum computing.
With rivals reporting 800x error reduction, can IBM's architectural fix for superconducting qubits truly close the quantum computing fidelity gap?
By solving the measurement bottleneck, what new challenge now becomes the main obstacle for fault-tolerant quantum computing?
Over 96% Logical Qubit Survival Achieved: A Major Leap Toward Fault-Tolerant Quantum Computing with IBM’s Heron r2
Overview
In June 2026, researchers from the University of Sydney and IBM achieved a major breakthrough by increasing logical qubit survival rates to over 96% on IBM’s 156-qubit Heron r2 superconducting processor. They tackled the problem of 'idling noise,' which causes errors during mid-circuit measurements, by redesigning quantum circuits and optimizing operation timing. This innovative engineering minimized the time qubits spent idle and exposed to noise, leading to a substantial reduction in error rates. The result marks a significant step toward more reliable and robust quantum computers, validated on advanced hardware with real experimental control.