Axi · Сквозной — фон контента (skin) · 16 кат.Axi · Сквозной — фон контента (skin) · 16 кат.

Google Willow Quantum Processor Surpasses Classical Strategy Limits

• Evgenia Sliv (edited: )
Google Willow Quantum Processor Surpasses Classical Strategy Limits

Researchers from Google Quantum AI demonstrated the advantage of the Willow quantum processor over classical strategies in several experimental tasks. The study results were published on November 19, 2026, in the scientific journal Science.

In the "magic square" game, the processor won 98.3% of rounds, with the maximum result of a classical strategy being 88.9%.

Scientists studied quantum contextuality – a property of quantum systems where the result of a measurement depends on which other compatible measurements are performed simultaneously. Researchers consider this phenomenon as one of the possible mechanisms providing the advantage of quantum algorithms over classical ones.

Willow won 98.3% of rounds in the quantum game

The experiment used the Mermin–Peres Magic Square game. Two participants are tasked with filling a 3 × 3 table row and column with numbers +1 and −1.

According to the game rules, the product of the numbers in the row must equal +1, and in the column −1. Moreover, the values at the intersection of the row and column must match. After receiving the tasks, participants cannot exchange information.

The best classical strategy allows winning no more than eight out of nine rounds, or approximately 88.9%. The use of quantum entanglement theoretically allows achieving 100%.

The Google team implemented the experiment on the Willow processor using four qubits forming two entangled pairs. Over 900,000 rounds, the system achieved a success rate of 98.3%.

Additionally, researchers tested the violation of the Bell–Kochen–Specker inequality, which is used to study quantum contextuality.

The measured indicator was 5.59, with a classical limit of 4 and a theoretical quantum maximum of 6. Thus, the processor demonstrated a result unattainable by classical strategy under the given conditions.

Experiments expanded to 105 qubits

In the next stage, researchers tested whether the advantage of quantum strategies is maintained as the system scale increases.

For this, a game based on Greenberger–Horne–Zeilinger (GHZ) states was used. In it, several participants must respond to tasks in a coordinated manner without exchanging information.

The team gradually increased the number of participants to 71, using one qubit per participant. According to Google Quantum AI results, the processor maintained an advantage over classical strategies at all tested scales.

However, as the number of qubits increased, the success rate decreased due to the accumulation of quantum operation errors.

Separately, scientists investigated the task of finding a Hidden Linear Function (HLF). For this, shallow quantum circuits on a two-dimensional qubit lattice were used.

During the experiment, the system scale was increased to 105 qubits. With the maximum number of qubits, the probability of a correct answer was about 60%.

Contextuality may help more accurately assess quantum processors

In addition to demonstrating quantum advantage, the study offers an additional way to test the quality of computational equipment.

Standard tests with random quantum circuits do not always detect small systematic errors. Experiments on contextuality may be more sensitive to such deviations, as they require the coordinated work of several qubits.

Games with mathematically established limits of classical strategies allow assessing the quality of quantum equipment without directly comparing computation speed with conventional computers.

However, the study results do not imply that Willow can perform any computational tasks faster than the best classical supercomputers. The demonstrated advantage relates to specific algorithms and experimental conditions.

Nevertheless, the work of Google Quantum AI shows that quantum contextuality persists in systems with a large number of qubits and can be used to develop new methods for evaluating quantum processors.

This material is prepared solely for informational purposes and does not constitute financial advice or a recommendation.

Comments

Checking authorization...
Loading comments...