A study published in Nature Physics on October 9, 2026, reported a dynamical transition between quantum and classical behavior in adaptive monitored circuits run on IBM hardware. The paper, “Order from chaos with adaptive circuits on quantum hardware”, studied circuits with up to 100 qubits. At L = 100 and p = 0.5, the experiment used about 5,000 entangling two-qubit gates and about 5,000 conditional resets—two distinct kinds of operation.

What the study observed

The researchers implemented a quantum version of the Bernoulli map, a classically chaotic system. Entangling operations scramble information; local measurements during the circuit and conditional resets feed information back into its evolution, steering it toward a fixed point. The study reported a dynamical transition between quantum and classical dynamics as this adaptive control varied.

The supplement tested system sizes from 10 to 100 qubits in steps of 10, and control probabilities from 0.2 to 0.85 in steps of 0.05. One finite-size fit over p = 0.4–0.6 estimated a critical value of pc = 0.4970(1). That estimate belongs to the stated fit and model.

What the operation counts mean

At the largest studied size, L = 100, and p = 0.5, the circuit used approximately 5,000 entangling two-qubit gates interspersed with approximately 5,000 measurement-induced conditional resets. The counts refer to separate operations under those settings. At L = 100 and p = 0.2, the supplement reports a maximum of about 8,000 two-qubit gates.

For each system-size and control-probability setting, the methods describe 50 circuit realizations, each repeated for 1,000 shots. The operation totals therefore belong to a particular experimental setup, not a general performance specification for the processor.

The processor behind the experiment

The study’s supplement identifies the backend as ibm_fez. IBM’s Quantum Platform lists ibm_fez as a Heron r2 processor with 156 programmable qubits. The experiment’s largest circuit used 100 qubits; the backend’s listed capacity is a separate hardware specification.

The paper’s result concerns adaptive control of quantum dynamics and the transition observed in that system. It is not a demonstration of a fully fault-tolerant quantum computer or general-purpose quantum error correction.