Breakthrough in quantum correction: logical qubit survival rate reaches 96% on IBM Heron
Quantum computing is taking another decisive step toward practical implementation. In a joint experiment conducted using the advanced 156-qubit superconducting processor IBM Quantum Heron r2, my team and partners from the University of Sydney achieved an increase in logical qubit preservation to 96% per error correction cycle. This is a significant improvement over previous figures, which did not exceed 90%.
The main stumbling block on the path to fault-tolerant quantum computing (FTQC) remains the so-called "idle noise." The essence of the problem is that to correct errors, the system is forced to regularly perform internal measurements of qubits. During these pauses, the remaining components of the processor lose stability, generating new errors. This creates a vicious cycle where error correction itself becomes a source of instability.
To break this vicious cycle, we completely redesigned the architecture of the correction circuits. The main goal was to radically reduce the time of forced computational idle. Optimizing the algorithms allowed not only to reduce pause times but also to significantly increase the system's resilience to emerging interference. The result on IBM Heron r2 is impressive: the survival rate of logical qubits jumped from less than 90% to 96%.
The project leader emphasized that this process is not a single event but a repeatedly occurring cycle at each stage of computation. Each such forced idle is a serious obstacle to reliable operation. Although the result was obtained in laboratory conditions on a single processor, the direction of research is critically important for the entire industry.
My analysis of the situation
Although 96% is an impressive laboratory result, commercial operation is still far off. The main challenges—scalability and fault tolerance—remain unresolved. However, IBM's progress in error correction and the company's plans to achieve quantum advantage by the end of 2026 now look much more realistic. If the pace of improvement continues, we could witness the beginning of a new era in computing within the next couple of years.