From quantum utility today to quantum advantage tomorrow: incumbent technology companies – among them Google, Amazon, IBM and Microsoft – and a wave of ambitious start-ups are on a mission to transform quantum computing from applied research endeavour to mainstream commercial opportunity. The end-game: quantum computers that can be deployed at-scale to perform computations significantly faster than classical machines while addressing scientific, industrial and commercial problems beyond the reach of today’s high-performance computing systems.
Meanwhile, as technology translation gathers pace across the quantum supply chain, government laboratories and academic scientists must maintain their focus on the “hard yards” of precompetitive research. That means prioritizing foundational quantum hardware and software technologies, underpinned by theoretical understanding, experimental systems, device design and fabrication – and pushing out along all these R&D pathways simultaneously.
Bringing order to disorder
Equally important is the requirement to understand and quantify the relative performance of quantum computers from different manufacturers as well as across the myriad platform technologies – among them superconducting circuits, trapped ions, neutral atoms as well as photonic and semiconductor processors. A case study in this regard is a broad-scope UK research collaboration that, for the past four years, has been reviewing, collecting and organizing a holistic taxonomy of metrics and benchmarks to evaluate the performance of quantum computers against their classical counterparts as well as the relative performance of competing quantum platforms.
Funded by the National Quantum Computing Centre (NQCC), which is part of the UK National Quantum Technologies Programme (NQTP), and led by scientists at the National Physical Laboratory (NPL), the UK’s National Measurement Institute, the cross-disciplinary consortium has taken on an endeavour that is as sprawling as it is complex. The challenge lies in the diversity of quantum hardware platforms in the mix; also the emergence of two different approaches to quantum computing – one being a gate-based framework for universal quantum computation, the other an analogue approach tailored to outperforming classical computers on specific tasks.
“Given the ambition of this undertaking, we tapped into a deep pool of specialist domain knowledge and expertise provided by university colleagues at Edinburgh, Durham, Warwick and several other centres-of-excellence in quantum,” explains Ivan Rungger, a principal scientist at NPL, professor in computer science at Royal Holloway, University of London, and lead scientist on the quantum benchmarking project. That core group consulted widely within the research community and with quantum technology companies across the nascent supply chain. “The resulting study,” adds Rungger, “positions transparent and objective benchmarking as a critical enabler for trust, comparability and commercial adoption of quantum technologies, aligning closely with NPL’s mission in quantum metrology and standards.”
Not all metrics are equal – or mature
’’For context, a number of performance metrics used to benchmark classical computers can also be applied directly to quantum computers, such as the speed of operations, the number of processing units, as well as the probability of errors to occur in the computation. That only goes so far, though, with all manner of dedicated metrics emerging in the past decade to benchmark the performance of quantum computers – ranging from their individual hardware components to entire applications.
Complexity reigns, it seems, and navigating the extensive literature can prove overwhelming, while the levels of maturity for different metrics varies significantly. Objective comparisons aren’t straightforward either – not least because variations of the same metric are commonly deployed; also the data disclosed together with a reported metric value is often not sufficient to reproduce the results.
“Many of the approaches provide similar overall qualitative performance values,” Rungger notes, “but the divergence in the technical implementation makes quantitative comparisons difficult and, by extension, slows progress of the field towards quantum advantage.”
The task then is to rationalize the metrics used to evaluate the performance for a given quantum hardware platform to a minimal yet representative set agreed across manufacturers, algorithm developers and end-users. These benchmarks also need to follow some agreed common approaches to fairly and objectively evaluate quantum computers from different equipment vendors.
With these objectives in mind, Rungger and colleagues conducted a deep-dive review that has yielded a comprehensive collection of metrics and benchmarks to allow holistic comparisons of quantum computers, assessing the quality of hardware components all the way to system-level performance and application-level metrics.
Drill down further and there’s a consistent format for each metric that includes its definition, a description of the methodology, the main assumptions and limitations, and a linked open-source software package implementing the methodology. The software transparently demonstrates the methodology and can also be used in practical, reproducible evaluations of all metrics.
“As research on metrics and benchmarks progresses, our collection of metrics and the associated software for performance evaluation are expected to evolve,” says Rungger. “Ultimately, the repository we have put together will provide a ‘living’ online resource, updated at regular intervals to account for community-driven developments in the field.”
From benchmarking to standards
Innovation being what it is, those developments are well under way. For starters, the importance of objective and relevant performance benchmarks for quantum computers has led several international standards bodies to initiate work on specific areas that are ready for standardization – work that, in turn, will give manufacturers, end-users and investors an informed evaluation of the performance of a range of quantum computing components, subsystems and full-stack platforms.
What’s evident is that the UK’s voice on metrics and benchmarking is already informing the collective conversation around standards development. “The quantum computing community and international standardization bodies are adopting a number of concepts from our approach to benchmarking standards,” notes Deep Lall, a quantum scientist in Rungger’s team at NPL and lead author of the study. “I was invited to present our work to a number of international standardization meetings and scientific workshops, opening up widespread international engagement with our research and discussions with colleagues across the benchmarking community.”
He continues: “We want the UK effort on benchmarking and metrics to shape the broader international effort. The hope is that the collection of metrics we have pulled together, along with the associated open-source software provided to evaluate them, will guide the development of standardized benchmarks for quantum computers and speed up the progress of the field towards practical quantum advantage.”
That’s a view echoed – and amplified – by Cyrus Larijani, NPL’s head of quantum programme. “As we move into the next phase of NPL’s quantum strategy, the importance of evidence-based decision making becomes ever-more critical,” he concludes. “By grounding our strategic choices in robust measurement science and real-world data, we ensure that our innovations not only push the boundaries of quantum technology but also deliver meaningful impact across industry and society.”
Further reading
Deep Lall et al. 2025 A review and collection of metrics and benchmarks for quantum computers: definitions, methodologies and software https://arxiv.org/abs/2502.06717
The headline take from NQCC
Quantum computing technology has reached the stage where a number of methods for performance characterization are backed by a large body of real-world implementation and use, as well as by theoretical proofs. These mature benchmarking methods will benefit from commonly agreed-upon approaches that are the only way to fairly, unambiguously and objectively benchmark quantum computers from different manufacturers.
“Performance benchmarks are a fundamental enabler of technology innovation in quantum computing,” explains Konstantinos Georgopoulos, who heads up the NQCC’s quantum applications team and is responsible for the centre’s liaison with the NPL benchmarking consortium. “How do we understand performance? How do we compare capabilities? And, of course, what are the metrics that help us to do that? These are the leading questions we addressed through the course of this study.
”If the importance of benchmarking is a given, so too is collaboration and the need to bring research and industry stakeholders together from across the quantum ecosystem. “I think that’s what we achieved here,” says Georgopoulos. “The long list of institutions and experts who contributed their perspectives on quantum computing was crucial to the success of this project. What we’ve ended up with are better metrics, better benchmarks, and a better collective understanding to push forward with technology translation that aligns with end-user requirements across diverse industry settings.”
End note: NPL retains copyright on this article.
The post Performance metrics and benchmarks point the way to practical quantum advantage appeared first on Physics World.
