University of Sydney, IBM Lift Qubit Survival Above 96% on 156-Qubit Processor

Q: Is improving qubit survival to 96% the key to fault-tolerant quantum computing, or just another incremental step?

It is an important incremental step, not the single breakthrough that delivers fault-tolerant quantum computing. The Sydney-IBM result matters because it improves the survival of a logical qubit during each error-correction cycle from under 90% to above 96% by reducing noise caused by mid-circuit measurements and idle time. That is a meaningful gain on real hardware, and it identifies measurement noise as a major bottleneck. In practical terms, this helps show where current superconducting systems are losing reliability and how circuit redesign can recover some of it. But 96% survival per cycle still implies roughly a 4% failure rate at that stage, which is far above the error levels usually associated with scalable fault tolerance. The broader field is targeting physical operation error rates around 10^-3 or lower for many leading schemes, and some newer qLDPC-based roadmaps point toward about 0.1% physical error rates to make overhead manageable. Fault tolerance depends on staying comfortably below threshold, not just improving one benchmark. So this result is best understood as a strong engineering advance in a critical subsystem, not the decisive crossing of the fault-tolerance line. It reduces one of today’s dominant error sources and supports IBM’s longer-term roadmap, but practical fault-tolerant quantum computing will still require much better gate and measurement fidelities, faster decoding, more efficient codes such as qLDPC, and large-scale hardware integration.

3 articles · Updated · HPCwire · Jul 2

Logical qubit survival rose from below 90% to more than 96% per error-correction cycle after University of Sydney researchers and IBM redesigned circuitry on a 156-qubit IBM Heron r2 processor.
Mid-circuit measurements emerged as a major source of failure because other qubits must idle during each check, letting noise build up even as the system tries to correct errors.
The team found measurement noise is among the dominant limits on reliable quantum logic operations in current devices, giving engineers a quantitative target for improving hardware.
Published in Nature Communications, the work advances a 2024 Sydney-IBM collaboration funded by IARPA and points to a clearer path toward scalable, fault-tolerant quantum computing.

Sources

Center100%

HPCwire8h ago

University of Sydney, IBM Improve Quantum Computer Error Correction, Boost Qubit Survival

Financial Times8h ago

The quantum computing revolution is closer than you think

electronicsonline.net.au22h ago

New benchmarks for quantum error correction

4 Sources

Is improving qubit survival to 96% the key to fault-tolerant quantum computing, or just another incremental step?

As quantum computers become more reliable, which real-world problems will they actually solve first?

Over 96% Logical Qubit Survival Achieved on IBM Heron r2—Paving the Way for Practical Fault-Tolerant Quantum Computers

Overview

In July 2026, researchers from the University of Sydney and IBM achieved a major breakthrough by increasing the survival rate of logical qubits to over 96% per error-correction cycle on the IBM Heron r2 processor. This advance addresses one of quantum computing’s biggest challenges: keeping quantum information intact. The team used quantum error correction, which protects fragile qubits from environmental noise by spreading information across several physical qubits. By applying gauge theory for low-overhead fault tolerance, they tackled key limitations in quantum logic operations and clearly defined the performance needed for reliable error checks, moving quantum computing closer to practical use.

...