Google Quantum AI tests quantum contextuality on up to 105 qubits

Google Quantum AI tests quantum contextuality on up to 105 qubits

A smartphone exhibiting the Google search originate cloak.

Google Quantum AI says it has shown contextuality, a quantum property thought to strength some algorithms, on as much as 105 qubits, in a paper on the hide of Science. The checks also expose delicate hardware errors, but the paper does now not insist an unconditional advantage over classical computers.

Google Quantum AI says a quantum property thought to strength some algorithms nonetheless holds on as much as 105 qubits. The overview, published on Thursday on the hide of Science, examined the property on Google’s superconducting quantum processors.

The property is named contextuality. It formula a size’s consequence is now now not mounted upfront, but is depending on which other measurements are made alongside it. Till now, it had mostly been shown in programs of true a couple of particles.

In a single test, Google’s qubits won a two-player sport, the Mermin-Peres magic square, 98.3% of the time. The handiest capacity classical technique wins handiest 88.9% of the time. A greater job the usage of 105 qubits also beat classical strategies with exiguous resources.

Google says the checks double as a sensitive skill to accept delicate hardware errors that other benchmarks miss. It says this could likely likely again pave the skill for uses in chemistry and offers science, although the experiments didn’t simulate either.

The paper is cautious about its limits. It says there just isn’t this kind of thing as a unconditional proof that quantum computers beat classical ones normally, and its quantum edge in a single sport shrank as extra qubits joined. The work first seemed as a preprint in December 2025.

Error-free hardware stays the change’s substantial fair. Google has said it wants a priceless, error-corrected quantum laptop by 2029, and Microsoft, which unveiled its Majorana 2 chip in June, has residing the the same plot.

IBM agreed to buy HRL because it confirmed a chip that corrects its hold errors, and Nvidia has simulated a 104-qubit error-corrected machine. Merchants are piling in too, with Quantinuum raising $1.68B in its IPO in June.

Salvage the TNW newsletter

Salvage the biggest tech records for your inbox a week.

Learn More

spot_imgspot_img

Subscribe

Related articles

spot_imgspot_img