Global Consortium Unveils Quantum Optimization Benchmarking Library to Track Path to Quantum Advantage

A global research consortium led by IBM Quantum has launched the Quantum Optimization Benchmarking Library (QOBLIB), an open-source initiative to evaluate quantum, classical, and hybrid algorithms for solving NP-hard combinatorial optimization problems. The library establishes a standardized framework for benchmarking across ten problem classes.

The 'Intractable Decathlon' includes 1,264 specific instances of problems such as Market Split, Low-Autocorrelation Binary Sequences, and Steiner Tree Packing. These instances span from tens to tens of thousands of decision variables and become computationally hard for classical solvers at large scales.

The library's model-independent architecture allows researchers to formulate problem instances in Mixed-Integer Programming (MIP), Integer Linear Programming (ILP), or Quadratic Unconstrained Binary Optimization (QUBO) representations. A public web portal tracks community progress, featuring an interactive complexity-landscape visualization and a live registry of best-known classical and quantum solutions.

At launch, QOBLIB logged over 2,600 submissions across 24 contributing institutions, including industrial end-users, national supercomputing centers, academic groups, and commercial quantum software providers. The library serves as an open standard to evaluate quantum optimization algorithms as hardware scales.