Quantum Volume, CLOPS and Other Performance Metrics Compared

Updated 2026-10-11 | 4 min read | QUANTUM (QNT) community

Why one number is never enough

A car is not described by its top speed alone, and a quantum computer is not described by its qubit count alone. A chip with many noisy qubits can be less useful than a chip with fewer clean ones. Over the years, researchers have proposed several scorecards to capture quality, speed and size. This page walks through the main ones in plain English. For the wider benchmark picture, see the benchmarks guide.

Quantum volume: a quality score

Quantum volume was proposed by IBM researchers Andrew Cross, Lev Bishop, Sarah Sheldon, Paul Nation and Jay Gambetta in a 2018 paper. They describe it as a single number that summarizes a computer's performance, and it quantifies the largest random circuit of equal width and depth that the machine successfully implements. "Equal width and depth" means a square circuit: as many qubits as layers of gates. The number is a measurement, not a calculation from a spec sheet. It depends on gate errors, measurement errors, crosstalk and how well qubits can talk to each other. Better connectivity and good compiling software raise it too, which is why transpilation matters.

In the common IBM style version, scores are written as powers of two. If the largest passing square circuit uses n qubits, the quantum volume is 2 to the power n. A score of 2^10 therefore means a 10 qubit by 10 layer random circuit passed. Each extra step doubles the number, so records climb fast without the machine doubling in size.

The record so far

Quantinuum Ltd reported a quantum volume of 2^23, or 8,388,608, on its System Model H2 trapped-ion computer on May 12, 2025. It said this completed a 2020 pledge to raise the number tenfold each year. Quantum Computing Report, dated September 18, 2025, reported a later quantum volume of 2^25, or 33,554,432, on the same System Model H2 family, and Wikipedia's table lists it too. We did not find that figure on a Quantinuum page we could open, so treat it as reported. Quantinuum's May 2025 blog calls quantum volume "not gameable", which is a vendor opinion. See the company profile and trapped ion machines for why ions score well: every qubit can interact with every other.

CLOPS: a speed score

CLOPS stands for Circuit Layer Operations Per Second. It was introduced in a 2021 paper by Andrew Wack and colleagues at IBM, titled "Quality, Speed, and Scale." The paper frames three attributes: quality (measured by quantum volume), speed (measured by CLOPS) and scale (the number of qubits). It builds CLOPS on the quantum volume experiments and shows how both classical and quantum parts affect throughput. In plain words, CLOPS asks how quickly the whole system, including the control computers that feed it, can run many circuits. That matters for hybrid methods such as variational circuits, which call the machine thousands of times.

Other metrics you will see

Wikipedia's quantum volume page also lists related ideas: cross entropy benchmarking, rQOPS (reliable quantum operations per second, proposed by Microsoft), CLOPS (IBM) and algorithmic qubits (proposed by IonQ). Each company tends to favor the metric that flatters its design. A table helps.

• Qubit count: scale only. Says nothing about error.
• Quantum volume: quality of a random square circuit. One number, one circuit family.
• CLOPS: speed of running layered circuits end to end.
• Algorithmic qubits: a vendor proposal that tries to count qubits usable on real algorithms.
• rQOPS: aimed at the error corrected era, counting reliable operations per second.
• Gate fidelity and T1 and T2 times: low level building blocks that feed all the scores above.

Known limits of quantum volume

Quantum volume tests only random square circuits. It does not show how a machine does on chemistry or optimization. Under the original definition, adding qubits can lower the score if depth does not keep up. A related family called volumetric benchmarks uses rectangular circuits to show time and space trade-offs, at the cost of a single clean figure. As error correction arrives, other measures may matter more. Read the NISQ guide and the error correction guide for that shift.

How to use these numbers

• Ask which circuits were run and who checked the result.
• Compare quality, speed and scale together, not one by one.
• Remember that a higher score signals potential, not delivered real world results.

The QNT memecoin is independent of Quantinuum Ltd and of any company named here. This page is education, not financial advice, and it predicts no token price.

Sources and further reading

• arXiv 1811.12926: Validating quantum computers using randomized model circuits (Cross et al.)
• arXiv 2110.14108: Quality, Speed, and Scale (Wack et al.)
• Quantinuum: quantum volume milestone blog
• Quantum Computing Report: Quantinuum reaches quantum volume 2^25 (Sept. 18, 2025)
• Wikipedia: Quantum volume

Reported as of 2026-10-09. Research moves fast, so check the original papers and company pages.

Frequently asked questions

Share on X

Keep reading

• Quantum Benchmarks Explained: Why Qubit Count Misleads
  How do you measure a quantum computer? Learn what error rates, fidelity, coherence time and quantum volume mean, and why qubit count alone is not enough.
• Decoherence and Noise Explained: What T1 and T2 Mean
  Why qubits lose their quantum behavior. A plain guide to decoherence, T1 energy relaxation and T2 dephasing, and why they limit quantum computers.
• Quantum Gates and Circuits Explained
  What are quantum gates and circuits? A plain English guide to Hadamard, CNOT and how gates turn qubits into a working quantum program.
• NISQ Explained: Noisy Intermediate-Scale Quantum Computers
  What does NISQ mean? Learn why today's noisy, mid-size quantum computers are limited, what they can do, and how the field plans to move past them.

All Quantum computing guides | Back to top | Search the site

Main pages: Quantum computing explained | Quantum and crypto | Companies | Quantum news | Glossary