Algorithmic qubits (#AQ) is a benchmark metric introduced by IonQ that attempts to quantify the number of "useful" qubits available for running real algorithms, after accounting for gate errors, connectivity constraints, and compilation overhead. Mathematically, #AQ is defined as the base-2 logarithm of the quantum volume: if a system has quantum volume 2^m, it has m algorithmic qubits. This reframing makes it easier to compare processors on a linear scale rather than the exponential quantum volume scale.
The motivation behind algorithmic qubits is to provide a more intuitive metric than quantum volume. Saying a processor has "25 algorithmic qubits" is easier to interpret than saying it has "quantum volume 33,554,432." It also aligns with the common question practitioners ask: "How many useful qubits do I have for my algorithm?" IonQ's roadmap targets 64 algorithmic qubits (QV = 2^64) as a key milestone for quantum advantage on commercially relevant problems.
Critics argue that algorithmic qubits inherit all the limitations of quantum volume (testing random circuits rather than structured algorithms, not capturing error correction capabilities) while adding a layer of rebranding that can obscure rather than illuminate. The metric is primarily used by IonQ and has not been widely adopted by other quantum computing companies. Nevertheless, the underlying concept — that the number of qubits is less important than the number of qubits that can perform useful work — reflects an important shift in how the industry evaluates quantum hardware.