MQT Bench 2.0 Released — Modular, Extensible Quantum Circuit Benchmarking
MQT Bench 2.0 brings a major overhaul to quantum circuit benchmarking — more modular, standards-aligned, and extensible, with new benchmarks, broader device support, and a cleaner API throughout.
MQT Bench 2.0 is a significant update to one of MQT’s most widely used tools for evaluating quantum circuits across abstraction levels and devices. The release focuses on making the benchmarking workflow more flexible and better integrated with the broader quantum software ecosystem.
Highlights of the new release include native compatibility with Qiskit’s Target class, support for bringing custom circuits into the benchmarking pipeline, and new benchmarks for arithmetic and variational algorithms — including contributions from BMW Group’s QUARK framework. The release also adds compilation to Clifford+T gates, mirror circuit generation for scalable hardware benchmarking, and export to OpenQASM 3 and QPY formats.
Under the hood, the API has been overhauled and documentation significantly improved, making MQT Bench easier to adopt and extend for both researchers and developers.
MQT Bench is fully open source. A redesigned web interface is also in the works.