IBM has announced three new experiments which it says demonstrate a “quantum advantage”: quantum processors produced results beyond practical classical simulation while using methods intended to show the answers can be trusted. This matters because previous claims have often been weakened by improved classical algorithms, and because today’s error-prone quantum machines can produce outputs too complex for conventional computers to verify directly.

One project, involving IBM, Japan’s RIKEN and error-mitigation firm Qedma, modelled a Floquet quantum process using an Ising model, in which a grid of interacting magnets changes over time. Researchers used creative error-handling and validation approaches rather than relying solely on smaller, classically checkable versions of the same calculation. The results are not immediately commercially useful, but IBM argues they provide stronger evidence that current hardware is progressing towards reliable quantum computation.