Superconducting Qubit Gate Sets Speed Record at 6.8 Nanoseconds With 99.98 Percent Accuracy

A team of physicists has pushed a superconducting quantum bit to perform a computational operation in 6.8 nanoseconds—roughly the time it takes light to travel two meters in a vacuum. The operation achieved an accuracy exceeding 99.98 percent, setting a new benchmark for gate speed. The advance addresses a critical bottleneck in quantum hardware development.[2][4]

The research, published October 23 in Nature Communications, was conducted on a single transmon qubit fabricated at the Helmholtz Nano Facility. The team from Forschungszentrum Jülich, RWTH Aachen University, and the University of Cologne tackled a fundamental trade-off. Historically, researchers have had to choose between executing fast gates or maintaining low error rates.[1][2]

"The performance of a quantum computer depends on how many computational steps it can carry out while keeping errors to a minimum," said Prof. Frank Wilhelm-Mauch of Forschungszentrum Jülich. He coordinates the QSolid project, which backed the research. Speeding up gate operations allows more steps to fit within a qubit's limited coherence window.[2]

To achieve the 6.8-nanosecond mark, the researchers had to solve the problem of quantum leakage. Superconducting transmons possess multiple energy levels beyond the binary zero and one used for computation. When driven by the strong, rapid microwave pulses required for fast operations, the qubit can accidentally absorb multiple photons.[1][4]

Suppressing unwanted transitions

This multi-photon absorption pushes the qubit into higher, unwanted energy states, effectively removing it from the calculation. The team developed a microwave pulse-shaping method called R2D to counter this effect. The technique applies an established control protocol, known as Derivative Removal by Adiabatic Gate (DRAG), twice recursively to the microwave signal.[1][2]

This double application specifically suppresses the frequency components that trigger multi-photon transitions. By shaping the pulse with exceptional precision, the R2D method reduces the probability of the qubit escaping its computational subspace by a factor of more than 20 compared to standard single-photon correction techniques.[1][4]

The resulting leakage error remained consistently below 2.0 × 10⁻⁵ per operation. To verify this figure, the researchers introduced an Amplified Leakage Error sequence. This new benchmarking tool coherently accumulates leakage errors, allowing the team to measure rates down to one in a million—exposing flaws that standard randomized benchmarking leaves buried in noise.[1][5]

While the speed and fidelity metrics represent a significant leap, analysts note the demonstration remains a proof of principle. The 99.981 percent and 99.989 percent accuracies were recorded for specific single-qubit gates—the X and X/2 operations—on an isolated device, rather than within a fully functioning multi-qubit processor.[2]

Shortening a single operation only benefits a larger computation if the control remains precise when surrounded by active neighbors. Simultaneous operations and two-qubit entanglement gates introduce complex crosstalk and interference. An isolated single-qubit test cannot capture the full error budget that emerges when multiple components interact.

The path to scaling up

The Jülich team acknowledged these limits, identifying tests on larger quantum processors as the necessary next step. The QSolid consortium, launched in early 2022, is currently evaluating a 30-qubit system. That larger platform will eventually reveal whether the R2D pulse-shaping method maintains its record-breaking performance under the demanding conditions of a scaled-up quantum computer.[2]

The drive for faster gates stems from the fragile nature of quantum information. Qubits lose their state rapidly due to environmental noise, a process called decoherence. If a system can execute operations in 6.8 nanoseconds rather than the typical tens of nanoseconds, it can complete deeper algorithms before the information degrades.[2][4]

This speed advantage becomes particularly crucial for quantum error correction. Error-correcting codes require continuous, rapid measurement and manipulation of physical qubits to protect the logical information they encode. If the physical gates are too slow, errors accumulate faster than the system can identify and correct them.[1]

"In QSolid, we have made important methodological advances, particularly in gate speed, that extend far beyond the project itself," Wilhelm-Mauch said. He noted that the results demonstrate highly capable components and show that the integration of hardware and software has reached a mature level.[2]

The researchers emphasized that the R2D pulse-shaping approach is not limited to superconducting transmons. Because the method relies on modifying the microwave control signals rather than altering the physical hardware, it can be adapted to other quantum computing platforms that suffer from similar multi-level leakage issues.[1][5]

Benchmarking beyond the noise

The paper uses randomized benchmarking to estimate total gate error, leakage benchmarking to track population outside the computational states, and purity benchmarking to separate decoherence contributions. Keeping those quantities distinct matters, as a leakage probability is not interchangeable with gate infidelity.

The Amplified Leakage Error sequence pairs repeated gate operations with a phase-matched virtual-Z rotation. This allows researchers to distinguish coherent gate leakage from thermal transitions, a distinction that standard methods struggle to make when precision drops below 10⁻⁵.[1]

As the quantum industry pushes toward fault-tolerant systems, techniques that isolate and eliminate specific error channels will define the next generation of hardware. The true test for the 6.8-nanosecond gate will arrive when it is deployed across dozens of interconnected qubits, facing the noise of a complete quantum circuit.

The next milestone for the Jülich team will involve reporting the full error budget alongside the timing for two-qubit operations. That integration will show exactly how much of the single-qubit speed gain survives when the hardware is tasked with running complex, multi-layered quantum algorithms.