Google has reported a significant step in quantum computing: its Willow processor reduced errors as the number of qubits used for error correction increased. The result, published alongside the company’s December 2024 announcement, addresses one of the field’s persistent engineering problems: quantum information is fragile, and adding more hardware can add more opportunities for errors.
What Google measured
Willow is a 105-qubit processor built from superconducting qubits. In the research, Google arranged qubits into increasingly large error-correcting codes and found that the encoded error rate fell as the code grew. The team described this as operating below the error-correction threshold: a crucial condition for scaling up, because adding qubits must eventually make a logical qubit more reliable rather than less.
That result is different from simply making a processor with more physical qubits. Quantum error correction uses groups of physical qubits to represent and protect a smaller number of logical qubits. It consumes hardware, but without it, the noise in real devices can overwhelm a computation before it finishes. Willow’s reported improvement is evidence that this trade-off can begin to work in the right direction.
A striking benchmark, with limits
Google also ran a random circuit sampling test, a benchmark designed to be exceptionally difficult for classical computers to reproduce. The company said Willow completed the test in under five minutes, while estimating that a leading classical supercomputer would need an impractical amount of time to match the result.
The benchmark demonstrates a difference in computational performance on that particular task. It does not show that Willow can now outperform conventional computers on everyday workloads, or that it can solve useful scientific problems faster. Random circuit sampling is not a practical application; it is a way to test quantum hardware and compare a specific calculation with classical simulation.
Why the error result matters
For years, researchers have had to contend with a frustrating pattern: larger quantum systems could be more powerful in principle, but their extra noise made reliable computation harder. The below-threshold error correction result is encouraging because it suggests that carefully designed systems can improve as they scale. It does not remove the need for further work on qubit quality, control electronics, decoding, and the overhead required to build useful logical qubits.
Google’s announcement is best understood as progress on a demanding piece of the larger problem, not as a finished quantum computer. The company still needs to demonstrate that error-corrected operations can be sustained at much greater scale and applied to calculations with value beyond a benchmark. Willow makes the case that this path is technically plausible; turning that milestone into useful quantum computing remains a substantial engineering challenge.
