Quantum Universal Operations Performance System
Comment: If there are significant independent articles over the next 6 to 12 months discussing this in some details (some hundreds of words) then those can be added and the page submitted via AfC for review. Currently this is an announcement of work that has not been peer reviewed. Those go on press releases, not here. Ldm1954 (talk) 13:09, 16 November 2026 (UTC)
Quantum Universal Operations Performance System (QUOPS) is an architecture-agnostic benchmarking framework introduced in September 2026 by researchers from Sandia National Laboratories, Quantinuum, and NVIDIA to measure and compare the real-world computational power of quantum computers. While other performance metrics such as Quantum Volume and CLOPS provide metrics for depth and speed, QUOPS aims to evaluate full-system execution. Doing this accounts for other properties a quantum computer implementation has that can affect calculations, such as compilation overhead, error rates, and decoding bottlenecks.[1]
The Calculation Process
- Construction of the Benchmark Circuit: Each circuit has a width w (the number of qubits) and a depth d (the number of gate layers). This gives a size s, defined as 2*width*depth. Physically, s is the total number of single-qubit rotations plus twice the number of CNOT gates.[2] This means that a single qubit rotation is 1 operation while a two-qubit gate operation counts as 2.
- Success and Accuracy Verification: To account for the fact that larger quantum circuits can’t be simulated on classical computers, the Mirror Circuit Fidelity Estimation is used. This estimation works by pairing a quantum circuit with its mirror image inverse, effective turning it into a self-checking loop. A circuit shape can only pass if its mean polarization reaches (about 61%) at 95% statistical confidence.[2]
- Mapping the Capability: All the passing circuit shapes in (w,s) form an shape known as the capability region. This is the boundaries of what the hardware can reliably execute. Two metrics are generated by running randomized circuits across different widths and depths:
- A QUOPS Score, denoted by Q measures the largest benchmark circuit size that can be ran inside something referred to as the utility cone, defined as . This essentially measures the largest standardized, computationally relevant circuit that a machine can successfully execute.[2]
- A QUOPS Rate, denoted by Ω measures what is effectively the operations per second. This is calculated after measuring Q and obtaining the total wall clock time τ (in seconds) with the squared process polarization γ̄², which acts to penalize the speed score depending on how error correction and post-selection are implemented. [2]
Together, the measurements “provide a two-dimensional view of capability while reducing system performance to a common currency.”[3] The two metrics form coordinates (Q, Ω) which can then be plotted graphically to compare architectures, track generational leaps, and forecast when quantum computers will be able to work on specific applications (chemistry simulations, factoring, material modeling, etc.). At its release in 2026, there was a 5-order-of-magnitude gap remaining between contemporary hardware QUOPs and utility-scale application requirements.
Table of Known Benchmarks
| Date | Manufacturer | Platform | Architecture | QUOPS (Q) | QUOPS/s (Ω) | Circuit Width |
| September 2026 | Willow | Physical | 216 | 20,000,000 | 6 | |
| September 2026 | IBM | Boston | Physical | 204 | 310,000 | 6 |
| September 2026 | Quantinuum | H2-1 | Physical | 1320 | 353 | 12 |
| September 2026 | Quantinuum | H2-1 | Physical with error detection | 1392 | 231 | 12 |
| September 2026 | Quantinuum | Helios-1 | Physical | 1504 | 303 | 16 |
| September 2026 | Quantinuum | Helios-1 | Physical with error detection | 1824 | 247 | 16 |
| September 2026 | Quantinuum | Helios-1 | [7,1,3] Code [[Steane code]] | 40 | 4.9 | 4 |
References
- ↑ Abdel-Kareem, Mohamed (2026-09-15). "Sandia, Quantinuum, and NVIDIA Introduce QUOPS Framework to Benchmark Physical and Logical Quantum Performance". Quantum Computing Report. Retrieved 2026-11-14.
- ↑ 2.0 2.1 2.2 2.3 Proctor, Timothy; Hart, Oliver; Maupin, Oliver Widzowski; Girling, Matthew; Hothem, Daniel; Mills, Daniel; Hines, Jordan; Mayer, Karl; Nelson, Jacob S.; LeBlond, Tyler; Chandani, Zohim; Forlivesi, Diego; Wysocki, Piper C.; Poór, Boldizsár; Dreiling, Joan M. (2026-09-10). "Benchmarking the computational power of quantum computers".
- ↑ "Introducing the Quantum Universal Operations Performance System: QUOPS". www.quantinuum.com. Retrieved 2026-11-14.
This article "Quantum Universal Operations Performance System" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:Quantum Universal Operations Performance System. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.
