Run PennyLane Circuits on Real Quantum Hardware
Use your existing PennyLane code, change one line, and run on real QPUs from IonQ, Rigetti, IQM, and AQT. Running in under a minute.
What is PennyLane?
PennyLane is Xanadu's open-source framework for differentiable quantum computing. It uniquely bridges quantum computing and machine learning by allowing you to train quantum circuits using automatic differentiation, just like training a neural network. This makes PennyLane the go-to framework for quantum machine learning (QML) and variational quantum algorithms.
With a rapidly growing community and a rich plugin ecosystem, PennyLane supports execution on simulators and real hardware through device plugins. Open Quantum provides a native PennyLane device plugin that integrates seamlessly with the qml.device interface, so your existing QNode functions and circuit code work without modification. Just swap the device and run on real QPUs.
How to Run PennyLane on Real Quantum Hardware
Install the Plugin
pip install openquantum-pennylane. Set your OPENQUANTUM_CLIENT_ID and OPENQUANTUM_CLIENT_SECRET environment variables. Done.
Write Circuits with PennyLane
Use standard PennyLane QNodes and operations. No vendor-specific code. No new syntax to learn. Your existing circuits work as-is.
Run on Any QPU
Pick a backend — IonQ, Rigetti, IQM, or AQT — and execute. Open Quantum handles transpilation, routing, and result formatting.
Your First PennyLane Circuit on a QPU in 60 Seconds
A complete example: pip install the plugin, create a Bell state with Python, and run it on IonQ Forte-1 trapped ion hardware. No credit card required.
pip install openquantum-pennylaneimport pennylane as qml
# Auth via OPENQUANTUM_CLIENT_ID and OPENQUANTUM_CLIENT_SECRET env vars
dev = qml.device("openquantum.device", wires=2, shots=1024, backend="ionq:forte-1")
@qml.qnode(dev)
def bell_state():
qml.Hadamard(0)
qml.CNOT(wires=[0, 1])
return qml.counts()
print(bell_state())Quantum Hardware Available for PennyLane
Run your PennyLane quantum machine learning circuits on trapped ion and superconducting QPUs from IonQ, Rigetti, IQM, and AQT. No vendor lock-in — switch backends with a single string change.
| QPU | Qubits | Technology | Highlights | Provider |
|---|---|---|---|---|
| IonQ Forte-1 | 36 qubits | Trapped Ion | 99.6% 2Q fidelity, software-configurable topology | Learn more |
| Rigetti Cepheus-1 | 108 qubits | Superconducting | 99.1% CZ fidelity, modular chiplet architecture | Learn more |
| IQM Emerald | 54 qubits | Superconducting | 99.5% CZ fidelity, European HPC integration | Learn more |
| IQM Garnet | 20 qubits | Superconducting | 99.5% CZ fidelity, European HPC integration | Learn more |
| AQT Ibex-Q1 | 12 qubits | Trapped Ion | Fully-connected, room temperature, rack-mounted | Learn more |