Skip to main content
Benchmarking quantum hardware is essential on the path toward practical quantum computing. As hardware advances, benchmarking must go beyond standard device-level tests such as randomized benchmarking or Quantum Volume to include algorithmic- and application-level evaluation, which measures what ultimately matters: the correctness and usefulness of the final computational outcome.

​
Functional-level benchmarks

Classiq provides a functional-level benchmarking suite for this purpose. Each benchmark is defined by a quantum model, a score, and a problem-size parameter, so you can study how performance scales with the task. Hardware constraints such as circuit width are handled at the synthesis stage rather than hard-coded into the benchmark, keeping the functional definition of a task separate from the capabilities of the target device.

​
Running across backends

Classiq exposes a wide variety of backends — QPUs, hardware emulators, and simulators across many providers — and lets you switch smoothly between them. The platform also handles the operational side of a run — managing execution budget (including running jobs against your Classiq-allocated budget, or emulating a device’s noise model without consuming QPU time or credits) and tracking the time each job takes to complete, so backends are directly comparable on both accuracy and turnaround.

​
Next steps

Run a set of predefined benchmarks covering canonical algorithmic and application-level tasks, or define your own. See Predefined Benchmarks to get started.