What is quantum advantage?

What the term means

The U.S. National Institute of Standards and Technology (NIST) describes quantum advantage as quantum computers outdoing classical computers at certain tasks. An older name is quantum supremacy. John Preskill used it in a 2012 paper for the day when well-controlled quantum systems can do things classical ones cannot.

In July 2025, IBM set two stricter conditions. The output must be rigorously validated, and the quantum approach must beat classical computation alone on efficiency, cost or accuracy.

Claims and counterclaims

By IEEE Spectrum's account, Google made the first claim, since contested. In a Nature paper published on October 23, 2019, its researchers said the 53-qubit Sycamore processor needed about 200 seconds to sample one quantum circuit a million times. They estimated the job would take a state-of-the-art supercomputer about 10,000 years. IBM researchers countered that month that a classical system could run an ideal simulation of the task in 2.5 days. They urged skepticism toward such claims.

NIST's explainer, updated in May 2026, says classical machines later matched or outperformed quantum processors on some tasks. It cites NIST physicist Scott Glancy as saying no early demonstration had proved truly useful.

In October 2025, Google Quantum AI and collaborators reported in Nature a measurement on 65-qubit circuits that took 2.1 hours per circuit. They estimated the same calculation would take about 3.2 years on the Frontier supercomputer, roughly 13,000 times longer. They said the results indicate a path to practical quantum advantage.

Where things stand in 2026

IBM's 2025 post predicted the research community would agree on the first demonstrations of advantage by the end of 2026. On July 30, 2026, IBM said three papers with researchers at the University of Chicago, Qedma and Algorithmiq demonstrated quantum advantage with built-in validation. The papers had not been peer reviewed, IEEE Spectrum reported. Dominik Hangleiter of ETH Zurich told the magazine all three were good science. But Hangleiter doubted the Qedma and Algorithmiq work showed a clear advantage, noting those papers do not claim one. Jens Eisert of the Free University of Berlin said establishing advantage is an ongoing process, not a single milestone.

Candidate claims are logged on the Quantum Advantage Tracker, whose contributors include IBM, university and national-laboratory researchers. It marks a claim superseded when later classical work closes or reverses the gap. On October 6, 2026, it listed six active and three superseded candidates.