Key Takeaways
- Using IBM’s 120-qubit Nighthawk r2 processor via standard cloud access, researchers produced 1 million random circuit samples in 19 seconds.
- Calculations via tensor-network contraction showed that Frontier, a leading exascale supercomputer, would need an estimated 110 years to finish the same workload.
- BlueQubit condensed matter physicist Tigran Sedrakyan and his team called it “the first demonstration of quantum advantage for a vanilla random-circuit sampling on a commercially and broadly accessible quantum processor that most non-expert quantum computer users can easily replicate.”
A milestone for accessible quantum power
In a demonstration of evolving computing capabilities, researchers have harnessed a commercially available quantum processor to complete a massive mathematical challenge in mere seconds. Running on IBM’s 120-qubit Nighthawk r2 hardware through standard cloud access, the system generated 1 million samples in just 19 seconds—a task estimated to demand roughly 110 years of computational effort from the fastest conventional supercomputers.
The benchmark relies on random circuit sampling, an operational framework long eyed by physicists as the ultimate test of quantum advantage over classical machines. Rather than operating within a secluded research facility with custom tuning, the experiment utilized standard platform workflows accessible to normal users.
“To our knowledge,” noted the research team led by theoretical condensed matter physicist Tigran Sedrakyan of BlueQubit, “this is the first demonstration of quantum advantage for a vanilla random-circuit sampling on a commercially and broadly accessible quantum processor that most non-expert quantum computer users can easily replicate.”
Finding the sweet spot in random circuit sampling
The race to prove practical quantum advantage gained significant momentum in 2018 when UC Berkeley theorists highlighted random circuit sampling as an ideal proving ground. In these experiments, operations are chosen at random to create an intricate, entangled quantum state across multiple qubits. While sampling the output state on a quantum machine requires measuring simple bit strings of ones and zeros, modeling the underlying probability distribution on classical hardware scales up exponentially in difficulty.
For this milestone, Sedrakyan and his colleagues engaged 61 qubits on the Nighthawk r2 processor. They tested progressively deeper circuits, reaching what they termed a Goldilocks zone at 36 cycles. At that depth, the system handled 918 two-qubit gates, generating the required 1 million samples in 19 seconds without requiring specialized hardware calibration.
Estimating the classical gap
Because world-class supercomputers are heavily booked and cannot simply be assigned century-long tasks, the researchers evaluated classical difficulty through tensor-network contraction. This mathematical method calculates the overall computational cost required to simulate quantum circuits on classical systems.
Their analysis revealed that recreating the million-sample distribution classically would require approximately 1.2 × 10²⁷ operations. When measured against the sustained performance of Frontier, the pioneering exascale system capable of more than a quintillion operations per second, completing the workload would require an estimated 110 years.
The researchers noted that this estimate does not represent an absolute biological or physical speed barrier for classical architectures, acknowledging that novel mathematical shortcuts and optimized algorithms regularly narrow such divides. However, by running on a “publicly available QPU accessible through a cloud,” the experiment underscores how accessible quantum platforms are becoming a viable reality for broader scientific exploration.
