Industry & Era
Algorithmic Qubits
A performance metric (popularized by IonQ) combining qubit count and gate fidelity to estimate how complex an algorithm a device can reliably run.
Definition
Algorithmic qubits is a metric intended to give a more meaningful picture of a quantum processor's practical capability than raw physical qubit count alone, by accounting for gate error rates and connectivity to estimate how large and complex a circuit the device can actually execute reliably.
In depth
Algorithmic qubits, abbreviated #AQ, is a metric created by IonQ to describe a quantum computer's usable size in a single number. It is based on a suite of application-oriented benchmarks developed by the Quantum Economic Development Consortium. A machine's #AQ is the largest number of qubits n for which it can run a set of algorithm circuits, such as quantum Fourier transform and amplitude estimation, whose depth grows with n, and still produce results that pass a fidelity threshold. A score of 36 means the machine handled the benchmark circuits up to a 36-qubit scale.
The metric was designed to reflect both qubit count and quality in one figure, since a chip with many noisy qubits can be less useful than one with fewer, better qubits. IonQ reported #AQ 25 for Aria in 2022, #AQ 36 for Forte in December 2024, and #AQ 64 for a Tempo development system in September 2025. The metric has been criticized for being defined and reported mainly by one vendor, for its dependence on the chosen benchmark circuits, and for being hard to compare with other metrics such as quantum volume or with the performance of superconducting machines.
Like all single-number metrics, #AQ simplifies. Real applications depend on specific circuit structures, connectivity, and error patterns. Industry efforts, such as the Quantum Economic Development Consortium's application benchmarks and the Metriq platform, try to provide transparent, reproducible comparison across vendors, and users should treat vendor-reported scores as one data point among several.
Technical Definition
The metric was introduced by IonQ as an alternative to simpler measures like Quantum Volume, intended to be easier to interpret directly in terms of practical algorithm capability rather than an abstract benchmark score.
Visual Explanation: An Analogy
Think of algorithmic qubits like a car's real-world fuel efficiency rating versus its theoretical maximum speed — raw qubit count is like top speed (easy to state, but not the full picture), while algorithmic qubits tries to capture how the device actually performs under realistic conditions.
Key facts
| Item | Detail |
|---|---|
| Defined by | IonQ, using QED-C application-oriented benchmarks |
| Aria | #AQ 25 (2022) |
| Forte | #AQ 36 (December 2024) |
| Tempo development system | #AQ 64 (September 2025) |
| Critique | Vendor-defined and dependent on the chosen benchmark circuits |
Where the idea comes from
IonQ introduced #AQ in 2020 and retired a related earlier measure. Competing metrics, such as IBM's quantum volume and circuit layer operations per second, address similar questions.
Real-World Use Cases
- Used by IonQ to argue that smaller, higher-fidelity trapped-ion systems can outperform larger, noisier superconducting processors for actual algorithm execution
- One of several competing metrics (alongside Quantum Volume) proposed for fairer cross-platform comparison
How it differs from related ideas
- Algorithmic qubits count usable qubits for algorithms, while quantum volume tests random square circuits.
- Neither equals the physical qubit count.
- A higher #AQ does not guarantee better performance on a specific application.
Frequently asked questions
What are algorithmic qubits?
IonQ's metric for the number of qubits a machine can use effectively to run a suite of application benchmarks.
How is #AQ different from the number of physical qubits?
It accounts for gate quality and connectivity, so it is usually lower than the physical qubit count.
Is #AQ the same as quantum volume?
No. They use different benchmark circuits and pass criteria, and values cannot be directly compared.
Related Terms
Keep exploring
- Algorithm: Quantum Fourier Transform (QFT)
- Algorithm: Quantum Amplitude Estimation
- Hardware: IonQ Forte (and Forte Enterprise)
- Hardware: IonQ Tempo