How to measure the performance of a quantum computer

IBM has introduced a new framework to evaluate quantum computing performance. As the industry moves toward more complex systems, the ability to measure progress accurately across different hardware designs becomes critical. IBM proposes focusing on three specific dimensions: scale, quality, and speed.

Scale is defined by programmable qubits, which represent the actual quantum bits users can control within an algorithm. This distinguishes usable resources from other physical components on a chip that support the hardware but do not perform direct calculations. By tracking this number, developers gain a clear view of the computational capacity available for their applications.

Quality is quantified by qubit operations. Currently, this focuses on two-qubit operations, which are the most challenging tasks a processor can reliably execute. Higher counts indicate that a system can run more complex circuits before errors impact the result. This metric serves as a practical gauge for the computational frontier of modern hardware.

Speed is measured by maximum circuits per second, often referred to as circuit throughput. This metric tracks how much useful work a system completes in a specific timeframe. High throughput reduces computational costs and allows users to run more tests efficiently. For instance, the upcoming Nighthawk r2 processor is expected to deliver significantly higher throughput than current fleets.

These metrics provide a consistent way to compare performance across various platforms, including superconducting circuits and trapped-ion systems. As the field advances toward fault-tolerant computing, the metrics will shift to reflect new priorities, such as logical qubits and T-gate counts. For now, however, these three indicators offer a transparent way to assess the utility and efficiency of existing quantum technology.