Sandia National Laboratories and Quantinuum have published peer-reviewed results in Nature demonstrating the performance of Quantinuum's 98-qubit Helios quantum computer, with single-qubit fidelity measured at 99.9975% and two-qubit fidelity at 99.921%. The figures make Helios the company's largest and highest-performing quantum system to date.

The paper, which appeared in Nature following an earlier arXiv posting, reports on the commercial system that debuted last year. The peer-reviewed publication means third-party experts have now scrutinized the findings — a distinction R&D managers should weigh when comparing vendor benchmark claims.

Sandia, home to the Department of Energy's longest-running quantum computing program, conducted the evaluation and certified the system's performance using a suite of tests, including benchmarking methods of its own invention. Notably, Sandia researchers supplied a new methodology to measure non-destructive readout operations — mid-circuit measurements — that are essential for quantum error correction. That detail matters for portfolio planning: mid-circuit measurement performance is a direct prerequisite for the fault-tolerant architectures DOE targets.

A four-year CRADA underpins the work

The assessment emerged from a Cooperative Research and Development Agreement between the two organizations, now in its fourth year and renewed in May. Sandia maintains similar agreements with several other quantum computing companies. The labs' quantum computing R&D program, running for more than 20 years, combines engineering, computer modeling, and microelectronics and nanotechnology facilities to build, characterize and share quantum devices across technology platforms.

The partnership formed a few years after Sandia researchers began developing integrated photonics for trapped-ion quantum computers — the architecture Helios uses — at the lab's Microsystems Engineering, Science and Applications complex. Integrated photonics are energy-efficient chips that carry information through microscopic optical channels at light speed; they aim to reduce energy costs and improve scalability, a key constraint on building large, useful machines. Sandia now helps Quantinuum design and test such components for possible inclusion in future platforms. Quantinuum, headquartered in Colorado, operates an R&D site in New Mexico near Sandia's main campus.

What the numbers do and don't say

The headline fidelities apply to one- and two-qubit operations. Helios operates beyond the capabilities of classical simulation alone, according to Tony Ransford, Helios lead architect at Quantinuum, and "established a new benchmark of fidelity and complexity for quantum computers."

Robin Blume-Kohout, a Sandia co-author on the paper, framed the results in terms of reliability rather than raw speed. "The most important aspect of today's quantum computers is not speed, but reliability," he said. Quantum computers can fail in dozens of subtle ways, he explained — an out-of-tune laser delivering bad instructions, a single atom jiggling out of place — and these failures degrade fidelity and cap performance. "We evaluate every aspect of quantum computer performance with our commercial partners to accelerate the advent of quantum supercomputing," he said.

Sandia senior manager Mike Descour positioned the collaboration as part of the laboratory's broader mandate. "As a national resource, we are committed to accelerating quantum computing technology in support of economic and national security," he said. Sandia assesses emerging opportunities and threats from quantum information science for the U.S. government across cryptography, pharmaceutical research, energy science, advanced sensing and communications.

The path forward

For labs and R&D organizations tracking quantum readiness, the results mark incremental but verified progress toward DOE's goal of fault-tolerant systems — machines large and reliable enough to solve complex problems. Photonics manager Chris DeRose signaled openness to widening the collaboration model: "We welcome collaboration with any interested partner including universities, industry and other national laboratories."

Blume-Kohout's long-term view frames the near-term milestone: helping companies solve reliability issues, he said, will bring about quantum computers capable of tackling unsolved scientific problems — and the Nature results show the nation reaching significant milestones along that path.