Title

A Reproducible Software Framework for Quantum Optimization Benchmarking

Abstract: 

Benchmarking quantum optimization methods requires more than comparing solver outputs or hardware execution times. Mathematical formulation, encoding, preprocessing, classical computation, and evaluation criteria can all affect the results.

We present a reproducible and solver-independent benchmarking framework based on JijModeling, OMMX, and Qamomile. The framework enables the same problem to be evaluated across classical, annealing, and gate-based quantum approaches. OMMX Quantum Benchmarks also provides benchmark instances, including those derived from QOBLIB.

Our framework supports consistent comparison of solution quality, feasibility, execution time, and computational resources while preserving the structure and semantics of the original problem.