Calibrating a superconducting transmon chip is a sequential decision problem under noise, drift, and a finite budget: an expert must choose experiments, read ambiguous plots, judge fit quality, and revise stale beliefs as the chip drifts. We study whether a vision-language agent can close this loop and specialize itself to one physical device without weight updates, via three co-designed artifacts. The first is a physics-grounded simulation environment for transmon chips: calibration observables derive from circuit-quantized parameters via scqubits, with realistic flux-line distortion, wall-time-scaled and mid-scan drift, and gate leakage, concerns a toy simulator would omit; each tool call advances a modeled clock so drift accrues by wall time, not call count. The second is a vision-language agent that runs the loop end to end, calling tools, reading plots, maintaining a structured notebook, and submitting parameters without hidden truth, scored against hidden parameters and gate fidelities measured on the device. The third is gradient-free online adaptation: a reflector reads truth-free anomaly signatures from past attempts and grows a small, human-readable device note appended to the prompt, admitted by a paired-snapshot accept gate that isolates strategy improvement from drift. On a hard-tier chip under budget pressure, six iterations raised the worst-case CZ fidelity from 0.678 to 0.787 and cut its variance, reproducing at four-qubit scale; a single accepted note raised CZ fidelity from 0.678 to 0.913 on its paired snapshot. A planted-fault study confirms the note is causal, diagnosing a hardware fault truth-free, its principal value raising the failure floor and cutting variance. The agent, scoring, and reward transfer to real hardware via a measurement-backend swap; only the accept gate is a simulation affordance, reducing to a held-out-slice or repeat-and-average form.
Calibrating a superconducting transmon chip is a sequential decision problem under noise, drift, and a finite budget: an expert must choose experiments, read ambiguous plots, judge fit quality, and revise stale beliefs as the chip drifts. We study whether a vision-language agent can close this loop and specialize itself to one physical device without weight updates, via three co-designed artifacts. The first is a physics-grounded simulation environment for transmon chips: calibration observables derive from circuit-quantized parameters via scqubits, with realistic flux-line distortion, wall-time-scaled and mid-scan drift, and gate leakage, concerns a toy simulator would omit; each tool call advances a modeled clock so drift accrues by wall time, not call count. The second is a vision-language agent that runs the loop end to end, calling tools, reading plots, maintaining a structured notebook, and submitting parameters without hidden truth, scored against hidden parameters and gate fidelities measured on the device. The third is gradient-free online adaptation: a reflector reads truth-free anomaly signatures from past attempts and grows a small, human-readable device note appended to the prompt, admitted by a paired-snapshot accept gate that isolates strategy improvement from drift. On a hard-tier chip under budget pressure, six iterations raised the worst-case CZ fidelity from 0.678 to 0.787 and cut its variance, reproducing at four-qubit scale; a single accepted note raised CZ fidelity from 0.678 to 0.913 on its paired snapshot. A planted-fault study confirms the note is causal, diagnosing a hardware fault truth-free, its principal value raising the failure floor and cutting variance. The agent, scoring, and reward transfer to real hardware via a measurement-backend swap; only the accept gate is a simulation affordance, reducing to a held-out-slice or repeat-and-average form.
Superconducting quantum computing is one of the most mature solid-state platforms for quantum computation, with processors exceeding one hundred qubits. Yet further scaling toward fault-tolerant quantum computing is increasingly constrained by calibration complexity. Conventional scripts are brittle to anomalous signals, and expert judgment is bounded by cognitive bandwidth and serial operation time, failing to keep pace with system scale. Here we report Vibe Calibration, an autonomous calibration system orchestrated by large language model agents, which distills expert tacit knowledge into reusable Skills. Each Skill is organized as a decision tree that packages parameterized measurement commands, quantitative acceptance criteria, and audit records, enabling autonomous execution and self-healing. We capture this knowledge through a three-phase human-in-the-loop distillation process and fine-tune a large language model on validated trajectories. On a 112-qubit processor with frequency-tunable transmons, the system autonomously completes calibration of 108 out of 112 qubits in 4.7 hours, achieving a 4--5× speedup over manual calibration of the full 112 qubits. A cross-validated comparison with expert manual calibration on a 16-qubit subset shows agreement on 14 out of 16 qubits. More importantly, the model demonstrates transferable calibration workflows across devices. While low-level control scripts require minor interface adaptation for different hardware platforms, the core decision logic and task orchestration generalize to new processors, demonstrating a reusable laboratory interface rather than a memorized script.This work demonstrates, for the first time, fully autonomous calibration of a hundred-qubit superconducting processor through reusable and auditable Skills, removing a critical barrier to scalable quantum hardware operation.
Quantum error correction imposes stringent requirements on the underlying hardware: all core operations of a quantum processor – single-qubit gates, two-qubit entangling gates, qubit readout, and qubit reset – must operate well below the error thresholds of leading codes. While such thresholds have already been crossed in several platforms, the associated overhead in physical qubits and circuit depth remains extremely sensitive to the physical error rate. This thesis investigates how far a superconducting platform based on transmon qubits can be pushed toward its physical limits, and how architectural and control choices can be optimized to reduce the overhead required for fault-tolerant quantum computation. First, we develop and characterize tunable-coupler architectures that enable fast, high-fidelity controlled-Z (CZ) gates. A long-distance coupler increases the qubit-qubit separation, freeing chip area for additional on-chip components, while still achieving CZ fidelities above 99.8%. Building on this, we optimize the qubit-coupler coupling to simultaneously support high-performance single- and two-qubit gates. We achieve simultaneous single- and two-qubit gate fidelities exceeding 99.9% on the same chip using a CZ calibration experiment based on our phase-averaged leakage error amplification (PALEA) protocol. PALEA coherently amplifies small over- and under-rotations of the diabatic CZ gate, making calibration errors easier to detect and correct. Second, we address single-qubit gates. Using spectrally engineered FAST-DRAG and higherderivative DRAG pulses, we demonstrate sub-8 ns gates with errors of order 10!" and strongly suppressed leakage. Third, we address readout by combining shelving to a higher qubit level with Purcell-filtered resonators and quantum-limited amplification, achieving readout errors below 10!#. Beyond the transmon workhorse, we explore the unimon qubit, which combines large anharmonicity with reduced charge sensitivity, and investigate a quantum circuit refrigerator as an integrated resource for fast, unconditional qubit reset via engineered dissipation. Finally, we demonstrate a cryogenic microwave link between two dilution refrigerators, enabling remote state transfer and entanglement over a five-meter distance and illustrating a path toward modular and distributed superconducting quantum processors. Together, these results show that with careful co-design of device layout, coupling topology, pulse shaping, and calibration protocols, all core operations – except qubit reset – of a superconducting quantum processor can be operated in the 10!#–10!" error regime within a coherent hardware and control stack, substantially lowering the resource overhead required for future fault-tolerant architectures.
As superconducting quantum processors scale toward hundreds of qubits, manual calibration becomes increasingly costly and difficult to sustain. Here we address two system-level factors governing automated calibration, namely, crosstalk induced by parallel operations and temporal drift of control parameters. On a 66-qubit superconducting processor, cross-entropy benchmarking shows that single-qubit calibration tolerates full-chip concurrency with minimal fidelity loss, whereas CZ calibration requires a minimum coupler-graph distance of four to avoid elevated control errors. Continuous monitoring of representative qubit and coupler parameters further yields quantitative refresh rules that assign refresh priority by error-budget consumption. These results turn two conventionally heuristic choices, the concurrency density of calibration tasks and the refresh priority of each control parameter, into measurable operating rules. Implemented within an automated calibration framework, the rules support a six-qubit Greenberger-Horne-Zeilinger state with a zero-noise-extrapolated fidelity of 89.64%, distance-dependent logical-error suppression in a single-cycle repetition-code benchmark, and a reduction of the mean single-qubit error from 0.023 to 0.011 on a 337-qubit processor, providing an experimentally grounded reference for the automated calibration and maintenance of current and future large-scale superconducting quantum systems.
Fault-tolerant quantum computing requires quantum gates that combine short duration, high fidelity, and robustness to noise. As quantum processors scale, robustness becomes increasingly important because exhaustive characterization and frequent recalibration of individual gates become impractical. We present a comprehensive optimal-control framework for controlled-Z (CZ) gates between directly coupled, frequency-tunable transmons that systematically accounts for intrinsic Hamiltonian errors and realistic decoherence mechanisms, including energy relaxation and dephasing. Using smooth, single-shot flux pulses, we obtain 12-ns CZ gates---close to the theoretical speed limit of 11.6 ns---with intrinsic errors suppressed to the 2×10−5 level. We construct dephasing noise models, calibrated against published experimental measurements, that capture fast dephasing from room-temperature control electronics and slow dephasing from 1/f flux noise. Using these models, we identify dephasing sweet spots that suppress sensitivity to dephasing noise. We further exploit transmon asymmetry as a design parameter to access distinct noise-resilient operating regimes. These results establish an experimentally applicable framework for balancing gate speed, fidelity, and noise robustness while reducing characterization and recalibration overhead.
The main source of error for superconducting quantum processors today is two-qubit gates, so understanding what limits their fidelity is central to improving these devices. This thesis studies the controlled-Z (CZ) gate between two transmon qubits, realized by tuning the ∣11⟩ and ∣20⟩ states through an avoided crossing with a flux pulse. There are two primary errors inherent to this approach. The first is leakage into non-computational states like ∣20⟩; the second is an exchange error from population transferring between the single-excitation states ∣01⟩ and ∣10⟩. Using a closed-system simulation, this work compares four flux-pulse shapes (unipolar, net-zero bipolar, Gaussian-filtered bipolar, and adiabatic) and constructs a CNOT from each CZ. The central question is which error channel sets the gate fidelity, and whether the widely used net-zero pulse, known to protect neighboring qubits, also helps inside the gate itself. An important finding of this research is that every type of flux pulse has an average gate fidelity limited by exchange error rather than leakage. While the exchange error is not apparent in the return of the ∣11⟩ population, because excitation-number conservation prevents ∣11⟩ from entering the single-excitation sector, it does appear as an off-diagonal term in the CZ unitary, and subsequently in the CNOT. A second finding concerns the net-zero pulse: although its zero net flux integral cancels the first-order phase imprinted on neighboring qubits, it does not suppress the exchange error. The strong suppression observed in the analytic bipolar gate happens because its idle time leaves the exchange oscillation near a cancellation point at the end of the gate, not because the pulse is net-zero. The flux operating point predicted by the simulation was tested on a physical transmon device accessed remotely, where a square pulse produced a 180∘ conditional phase close to the simulated flux and duration. Because the model is closed-system, the reported fidelities describe the coherent error of each pulse rather than the total infidelity of a physical device.
Superconducting transmon qubits are a leading technology for quantum information processing, yet their reliable operation rests on meticulous calibration and characterization routines. These processes have been fine-tuned and are relatively well understood by the quantum computing community. Nevertheless, it is often challenging for newcomers to compile all the available information into a practical experimental flow. In this tutorial, we present a comprehensive walkthrough for the characterization and optimization of tunable transmon qubits, demonstrated on a commercial five-qubit processor. Moving beyond theoretical description, we detail in a straightforward manner the complete workflow, from cryogenic setup and wiring to parametric amplifier optimum operation, flux sweet-spot identification, pulse calibration, and readout optimization. We also demonstrate the characterization of qubit-qubit coupling, covering all steps before multiqubit operations. This guide serves as a reference for experimentalists seeking to efficiently bring up transmon-based quantum devices.