Per-QPU Solver Properties and Schedules

The following sections provide information for advanced users who want to better understand and leverage the physical implementation of D-Wave’s various quantum processing units (QPUs) available in the Leap service[1]. This information includes:

• Summary of a QPU’s physical properties—The values provided are the physical properties of a calibrated QPU; they are not QPU specifications.

  Note

  In addition to the physical properties listed herein, each QPU has a number of other properties defined in software that are accessible via the Solver API. For a global list of the solver properties for a QPU, see the General QPU Solver Properties page and for a list of the permitted user parameters for each type of solver, see the QPU Solver Parameters page. To retrieve the solver properties for a particular QPU, see the examples on those pages.

• Spreadsheet for a QPU’s annealing-schedule functions and normalized annealing-waveform values—These values are required for computing the energy of a problem at a specific point in a QPU’s annealing process; as such, the spreadsheet provides the values to use for the \(A(s)\) and \(B(s)\) terms in the Hamiltonian of equation (1) for each value of the normalized anneal fraction \(s\), between 0 and 1 in increments of 0.001. Units for these terms are GHz, where the conversion from energy in Joules to Hz is through a division by Planck’s constant as follows:

  \[ \begin{align}\begin{aligned}A(s)_{\text{[GHz]}} &= \frac{A(s)_{\text{[Joules]}}} {6.62607004 \times 10^{-34} \times 10^9}\\&= 1.5092 \times 10^{24} A(s)_{\text{[Joules]}}\end{aligned}\end{align} \]

[1]

Depending on your customer contract, customer plan, and seat type in a project, you may not have access to all QPUs in the Leap service.

Advantage2_system1

All data presented in this section are specific to the Advantage2_system1 solver. The Advantage2™ system QPU is based on a physical lattice of qubits and couplers known as the Zephyr™ topology. For information, see the Zephyr Graph section.

Physical Properties

This table lists the physical properties of the calibrated QPU.

Table 24 QPU Physical Properties[2]

Property	Value
Model	\(\text{Advantage2}\)
Graph size	\(\text{Z12}\)
Number of qubits	\(4578\)
Number of couplers	\(41531\)
Qubit temperature	\(20 \pm 1.0\ \text{mK}\)
\(\rm M_{\rm AFM}\): Maximum mutual inductance for qubit pairs	\(0.528\ \text{pH}\)
Quantum critical point for 1D chains	\(2.308\ \text{GHz}\)
\(L_q\): Qubit inductance	\(120\ \text{pH}\)
\(C_q\): Qubit capacitance	\(147\ \text{fF}\)
\(I_c\): Qubit critical current	\(4.75\ \text{µA}\)
Average single-qubit temperature	\(0.112\)
Ferromagnetic-problem freezeout	\(0.008\)
Single-qubit freezeout	\(0.684\)
\(\Phi_{\rm CCJJ}^i\): Initial (at \(s=0\)) external flux on compound Josephson junctions	\(-0.694\ \Phi_0\)
\(\Phi_{\rm CCJJ}^f\): Final (at \(s=1\)) external flux on compound Josephson junctions	\(-0.760\ \Phi_0\)
Readout time range	\(17.0\ \text{to}\ 101.0\ \text{µs}\)
Programming time	\(\sim 33600\ \text{µs}\)
QPU-delay-time per sample	\(60.6\ \text{µs}\)
Readout error rate	\(\leq 0.001\)

[2]

Some notes for the QPU properties are as follows:

• The ferromagnetic problem and single-qubit freezeout points shown are normalized anneal fraction values.

• The average single-qubit temperature is given in units of linear bias, the dimensionless h of the Ising Hamiltonian.

• Readout time range: Typical readout times for reading between one qubit and the full QPU.

• Programming time: Typical for problems run on this QPU. Actual problem programming times may vary slightly depending on the nature of the problem.

• Readout error rate: Error rate when reading the full system.

Annealing Schedule

Download the annealing schedule for the QPU here: Advantage2_system1 Excel spreadsheet.

The standard annealing schedule for the QPU is shown in Figure 78.

Fig. 78 Standard annealing schedule for the QPU, showing energy changes as a function of scaled time.

Advantage2_system2.1

All data presented in this section are specific to the Advantage2_system2.1 solver. The Advantage2 system QPU is based on a physical lattice of qubits and couplers known as the Zephyr topology. For information, see the Zephyr Graph section.

Physical Properties

This table lists the physical properties of the calibrated QPU.

Table 25 QPU Physical Properties[3]

Property	Value
Model	\(\text{Advantage2}\)
Graph size	\(\text{Z12}\)
Number of qubits	\(4516\)
Number of couplers	\(40448\)
Qubit temperature	\(16.7 \pm 1.0\ \text{mK}\)
\(\rm M_{\rm AFM}\): Maximum mutual inductance for qubit pairs	\(0.554\ \text{pH}\)
Quantum critical point for 1D chains	\(1.853\ \text{GHz}\)
\(L_q\): Qubit inductance	\(140\ \text{pH}\)
\(C_q\): Qubit capacitance	\(173\ \text{fF}\)
\(I_c\): Qubit critical current	\(4.64\ \text{µA}\)
Average single-qubit temperature	\(0.119\)
Ferromagnetic-problem freezeout	\(0.023\)
Single-qubit freezeout	\(0.605\)
\(\Phi_{\rm CCJJ}^i\): Initial (at \(s=0\)) external flux on compound Josephson junctions	\(-0.668\ \Phi_0\)
\(\Phi_{\rm CCJJ}^f\): Final (at \(s=1\)) external flux on compound Josephson junctions	\(-0.735\ \Phi_0\)
Readout time range	\(17.0\ \text{to}\ 95.0\ \text{µs}\)
Programming time	\(\sim 24500\ \text{µs}\)
QPU-delay-time per sample	\(20.6\ \text{µs}\)
Readout error rate	\(\leq 0.001\)

[3]

Some notes for the QPU properties are as follows:

• The ferromagnetic problem and single-qubit freezeout points shown are normalized anneal fraction values.

• The average single-qubit temperature is given in units of linear bias, the dimensionless h of the Ising Hamiltonian.

• Readout time range: Typical readout times for reading between one qubit and the full QPU.

• Programming time: Typical for problems run on this QPU. Actual problem programming times may vary slightly depending on the nature of the problem.

• Readout error rate: Error rate when reading the full system.

Annealing Schedule

Download the annealing schedule for the QPU here: Advantage2_system2.1 Excel spreadsheet.

The standard annealing schedule for the QPU is shown in Figure 79.

Fig. 79 Standard annealing schedule for the QPU, showing energy changes as a function of scaled time.

Advantage2_system4.3

All data presented in this section are specific to the Advantage2_system4.3 solver. The Advantage2™ system QPU is based on a physical lattice of qubits and couplers known as the Zephyr™ topology. For information, see the Zephyr Graph section.

Physical Properties

This table lists the physical properties of the calibrated QPU.

Table 26 QPU Physical Properties[4]

Property	Value
Model	\(\text{Advantage2}\)
Graph size	\(\text{Z6}\)
Number of qubits	\(1203\)
Number of couplers	\(10553\)
Qubit temperature	\(18 \pm 1.0\ \text{mK}\)
\(\rm M_{\rm AFM}\): Maximum mutual inductance for qubit pairs	\(0.514\ \text{pH}\)
Quantum critical point for 1D chains	\(2.154\ \text{GHz}\)
\(L_q\): Qubit inductance	\(119\ \text{pH}\)
\(C_q\): Qubit capacitance	\(170\ \text{fF}\)
\(I_c\): Qubit critical current	\(4.96\ \text{µA}\)
Average single-qubit temperature	\(0.119\)
Ferromagnetic-problem freezeout	\(0.007\)
Single-qubit freezeout	\(0.671\)
\(\Phi_{\rm CCJJ}^i\): Initial (at \(s=0\)) external flux on compound Josephson junctions	\(-0.688\ \Phi_0\)
\(\Phi_{\rm CCJJ}^f\): Final (at \(s=1\)) external flux on compound Josephson junctions	\(-0.747\ \Phi_0\)
Readout time range	\(17.0\ \text{to}\ 45.0\ \text{µs}\)
Programming time	\(\sim 8000\ \text{µs}\)
QPU-delay-time per sample	\(20.6\ \text{µs}\)
Readout error rate	\(\leq 0.001\)

[4]

Some notes for the QPU properties are as follows:

• The ferromagnetic problem and single-qubit freezeout points shown are normalized anneal fraction values.

• The average single-qubit temperature is given in units of linear bias, the dimensionless h of the Ising Hamiltonian.

• Readout time range: Typical readout times for reading between one qubit and the full QPU.

• Programming time: Typical for problems run on this QPU. Actual problem programming times may vary slightly depending on the nature of the problem.

• Readout error rate: Error rate when reading the full system.

Annealing Schedule

Download the annealing schedule for the QPU here: Advantage2_system4.3 Excel spreadsheet.

The standard annealing schedule for the QPU is shown in Figure 80.

Fig. 80 Standard annealing schedule for the QPU, showing energy changes as a function of scaled time.

Advantage_system6.4

All data presented in this section are specific to the Advantage_system6.4 solver. The Advantage QPU is based on a physical lattice of qubits and couplers known as the Pegasus™ topology. For information, see the Pegasus Graph section.

Physical Characteristics

This table lists the physical properties of the calibrated QPU.

Table 27 QPU Physical Properties[5]

Property	Value
Model	\(\text{Advantage, performance update}\)
Graph size	\(\text{P16}\)
Number of qubits	\(5612\)
Number of couplers	\(40088\)
Qubit temperature	\(16.0 \pm 0.1\ \text{mK}\)
\(\rm M_{\rm AFM}\): Maximum mutual inductance for qubit pairs	\(1.554\ \text{pH}\)
Quantum critical point for 1D chains	\(1.281\ \text{GHz}\)
\(L_q\): Qubit inductance	\(382\ \text{pH}\)
\(C_q\): Qubit capacitance	\(119\ \text{fF}\)
\(I_c\): Qubit critical current	\(1.99\ \text{µA}\)
Average single-qubit temperature	\(0.221\)
Ferromagnetic-problem freezeout	\(0.073\)
Single-qubit freezeout	\(0.616\)
\(\Phi_{\rm CCJJ}^i\): Initial (at \(s=0\)) external flux on compound Josephson junctions	\(-0.624\ \Phi_0\)
\(\Phi_{\rm CCJJ}^f\): Final (at \(s=1\)) external flux on compound Josephson junctions	\(-0.723\ \Phi_0\)
Readout time range	\(18.0\ \text{to}\ 173.0\ \text{µs}\)
Programming time	\(\sim 14200\ \text{µs}\)
QPU-delay-time per sample	\(20.5\ \text{µs}\)
Readout error rate	\(\leq 0.001\)

[5]

Some notes for the QPU properties are as follows:

• Qubit temperature: Some qubits in this QPU are affected by high-frequency photon flux and may have a higher temperature than what is reported here. For more information, see the discussion of high-energy photon flux in the Errors and Error Correction section.

• The ferromagnetic problem and single-qubit freezeout points shown are normalized anneal fraction values.

• The average single-qubit temperature is given in units of linear bias, the dimensionless h of the Ising Hamiltonian.

• Readout time range: Typical readout times for reading between one qubit and the full QPU.

• Programming time: Typical for problems run on this QPU. Actual problem programming times may vary slightly depending on the nature of the problem.

• Readout error rate: Error rate when reading the full system.

Annealing Schedule

Download the annealing schedule for the QPU here: Advantage_system6.4 Excel spreadsheet.

The standard annealing schedule for this QPU is shown in Figure 81.

Fig. 81 Standard annealing schedule for the QPU, showing energy changes as a function of scaled time.

DAC Quantization Effects

The on-QPU digital-analog converters (DACs) that provide the user-specified \(h\) and \(J\) values have a finite quantization step size. That step size depends on the value of the \(h\) and \(J\) applied because the response to the DAC output is nonlinear.

Figure 82 and Figure 83 show the effects of the DAC quantization step for the DACs controlling the \(h\) and \(J\) values, respectively, for this system.

Fig. 82 Typical quantization on the \(h\) DAC control.

Fig. 83 Typical quantization on the \(J\) DAC control.

Advantage_system4.1

All data presented in this section are specific to the Advantage_system4.1 solver. The Advantage QPU is based on a physical lattice of qubits and couplers known as the Pegasus™ topology. For information, see the Pegasus Graph section.

Physical Properties

This table lists the physical properties of the calibrated QPU.

Table 28 QPU Physical Properties[6]

Property	Value
Model	\(\text{Advantage, performance update}\)
Graph size	\(\text{P16}\)
Number of qubits	\(5627\)
Number of couplers	\(40279\)
Qubit temperature	\(15.4 \pm 0.1\ \text{mK}\)
\(\rm M_{\rm AFM}\): Maximum mutual inductance for qubit pairs	\(1.647\ \text{pH}\)
Quantum critical point for 1D chains	\(1.391\ \text{GHz}\)
\(L_q\): Qubit inductance	\(372\ \text{pH}\)
\(C_q\): Qubit capacitance	\(119\ \text{fF}\)
\(I_c\): Qubit critical current	\(2.1\ \text{µA}\)
Average single-qubit temperature	\(0.198\)
Ferromagnetic-problem freezeout	\(0.064\)
Single-qubit freezeout	\(0.612\)
\(\Phi_{\rm CCJJ}^i\): Initial (at \(s=0\)) external flux on compound Josephson junctions	\(-0.621\ \Phi_0\)
\(\Phi_{\rm CCJJ}^f\): Final (at \(s=1\)) external flux on compound Josephson junctions	\(-0.717\ \Phi_0\)
Readout time range	\(17.0\ \text{to}\ 235.0\ \text{µs}\)
Programming time	\(\sim 14100\ \text{µs}\)
QPU-delay-time per sample	\(20.5\ \text{µs}\)
Readout error rate	\(\leq 0.001\)

[6]

Some notes for the QPU properties are as follows:

• Qubit temperature: Some qubits in this QPU are affected by high-frequency photon flux and may have a higher temperature than what is reported here. For more information, see the discussion of high-energy photon flux in the Errors and Error Correction section.

• The ferromagnetic problem and single-qubit freezeout points shown are normalized anneal fraction values.

• The average single-qubit temperature is given in units of linear bias, the dimensionless h of the Ising Hamiltonian.

• Readout time range: Typical readout times for reading between one qubit and the full QPU.

• Programming time: Typical for problems run on this QPU. Actual problem programming times may vary slightly depending on the nature of the problem.

• Readout error rate: Error rate when reading the full system.

Annealing Schedule

Download the annealing schedule for the QPU here: Advantage_system4.1.

The standard annealing schedule for this QPU is shown in Figure 84.

Fig. 84 Standard annealing schedule for the QPU, showing energy changes as a function of scaled time.

DAC Quantization Effects

The on-QPU digital-analog converters (DACs) that provide the user-specified \(h\) and \(J\) values have a finite quantization step size. That step size depends on the value of the \(h\) and \(J\) applied because the response to the DAC output is nonlinear.

Figure 85 and Figure 86 show the effects of the DAC quantization step for the DACs controlling the \(h\) and \(J\) values, respectively, for this system.

Fig. 85 Typical quantization on the \(h\) DAC control.

Fig. 86 Typical quantization on the \(J\) DAC control.

Advantage2_research1

All data presented in this section are specific to the Advantage2_research1 solver, which is based on a physical lattice of qubits and couplers known as the Zephyr™ topology. For information, see the Zephyr Graph section.

Physical Properties

This table lists the physical properties of the calibrated QPU.

Table 29 QPU Physical Properties[7]

Property	Value
Model	\(\text{Advantage2}\)
Graph size	\(\text{Z6}\)
Number of qubits	\(1176\)
Number of couplers	\(10170\)
Qubit temperature	\(17.5 \pm 1.0\ \text{mK}\)
\(\rm M_{\rm AFM}\): Maximum mutual inductance for qubit pairs	\(0.443\ \text{pH}\)
Quantum critical point for 1D chains	\(2.014\ \text{GHz}\)
\(L_q\): Qubit inductance	\(107\ \text{pH}\)
\(C_q\): Qubit capacitance	\(173\ \text{fF}\)
\(I_c\): Qubit critical current	\(4.57\ \text{µA}\)
Average single-qubit temperature	\(0.102\)
Ferromagnetic-problem freezeout	\(0.008\)
Single-qubit freezeout	\(0.603\)
\(\Phi_{\rm CCJJ}^i\): Initial (at \(s=0\)) external flux on compound Josephson junctions	\(-0.726\ \Phi_0\)
\(\Phi_{\rm CCJJ}^f\): Final (at \(s=1\)) external flux on compound Josephson junctions	\(-0.819\ \Phi_0\)
Readout time range	\(17.0\ \text{to}\ 87.0\ \text{µs}\)
Programming time	\(\sim 18200\ \text{µs}\)
QPU-delay-time per sample	\(20.6\ \text{µs}\)
Readout error rate	\(\leq 0.001\)

[7]

Some notes for the QPU properties are as follows:

• The ferromagnetic problem and single-qubit freezeout points shown are normalized anneal fraction values.

• The average single-qubit temperature is given in units of linear bias, the dimensionless h of the Ising Hamiltonian.

• Readout time range: Typical readout times for reading between one qubit and the full QPU.

• Programming time: Typical for problems run on this QPU. Actual problem programming times may vary slightly depending on the nature of the problem.

• Readout error rate: Error rate when reading the full system.

Annealing Schedule

Download the annealing schedule for the QPU here: Advantage2_research1 Excel spreadsheet.

The standard annealing schedule for the QPU is shown in Figure 87.

Fig. 87 Standard annealing schedule for the QPU, showing energy changes as a function of scaled time.

Working-Graph Changes

The Working-Graph Changes table identifies changes to the working graphs for the generally available QPU solvers. A QPU solver’s current graph_id is displayed on the Leap dashboard (a solver’s Properties and a problem’s Parameters modals) and available through the Ocean SDK.

Table 30 Working-Graph Changes

Solver Name[8]	Deployment Date[9]	Removed Qubits
Advantage2_system1
Advantage2_system1	2026-04-10	4374
Advantage2_system1.13	2026-03-18	249
Advantage2_system1.12	2026-03-04	1871, 1895
Advantage2_system1.11	2026-01-28	380, 381, 441, 577, 4202, 4226, 4237
Advantage2_system1.10	2025-12-15	2782
Advantage2_system1.9	2025-12-09	2049
Advantage2_system1.8	2025-11-19	484
Advantage2_system1.7	2025-11-04	510
Advantage2_system1.6	2025-09-02	3122
Advantage2_system1.5	2025-08-07	4058, 4082
Advantage2_system1.4	2025-07-16	769
Advantage2_system1.3	2025-06-20	2838
Advantage2_system1.2	2025-06-17	1283, 4686
Advantage2_system1.1	2025-05-20	First solver release
Advantage2_system4
Advantage2_system4.3	2025-12-22	840
Advantage2_system4.1	2025-11-03	First solver release
Advantage_system6
Advantage_system6.4	2024-02-28	104, 119
Advantage_system6.3	2023-10-18	Only couplers removed
Advantage_system6.2	2023-05-31	727, 742
Advantage_system6.1	2022-05-12	First solver release
Advantage_system4
Advantage_system4.1	2021-10-05	First solver release
Advantage2_research1
Advantage2_research1	2026-04-16	101, 604
Advantage2_research1.5	2025-12-18	Only couplers removed
Advantage2_research1.4	2025-11-26	5, 11, 29, 34, 41, 53, 58, 70, 89, 94, 106, 113, 118, 125, 130, 131, 1134, 1140, 1141, 1146, 1153, 1164
Advantage2_research1.3	2025-10-21	44, 466, 550, 622, 686, 735, 912, 1170, 1231
Advantage2_research1.2	2025-09-10	90, 434, 663, 693, 780, 868
Advantage2_research1.1	2025-08-14	First solver release

[8]

Since 2026-04-06, solvers that have their working graph updated no longer have a minor version in the solver name.

[9]

Deployment dates are links to the related release note, if available.