98.30% win rate Google Willow, struggles beyond upper limit in magic square game
Original headline98.30% 승률 구글 윌로우, 마방진 게임서 고전 상한 넘어
Not investment advice. Muhabbit scores how much a story may matter, not which way to trade.
Google Quantum AI's superconducting Willow processor achieved a 98.30% win rate in a four-qubit implementation of the Mermin-Peres magic square game, surpassing the 88.89% theoretical ceiling for classical strategies, the company disclosed in a research blog post on the 19th alongside a paper published as the cover article in Science the same day. The cooperative game assigns one player rows and the other columns of a 3×3 grid filled with +1 or -1 values, with no communication allowed, and requires every row product to equal +1 while every column product equals -1, a constraint that forces any pre-shared classical lookup table to fail on at least one of the nine combinations. The result matters as a benchmark demonstration that quantum hardware can outperform the best possible classical strategy in a communication-free setting, a step toward practical quantum advantage beyond narrow sampling tasks. What to watch next is whether Google extends the demonstration to larger qubit counts and whether the win rate holds as error rates and circuit depth grow.
Muhabbit's AI wrote this summary in its own words from TokenPost's report and translated it from Korean. It can get details wrong, so check the source.
Named in this story
Google Quantum AI, Willow, Science