EstimatorV2

menos.qiskit.MENOSEstimator implements Qiskit's BaseEstimatorV2. Given circuits and Pauli observables it returns exact expectation values.

MENOSEstimator(*, default_precision: float = 0.0, noise_model: NoiseModel | None = None)
MENOSEstimator.run(pubs, *, precision=None) -> PrimitiveJob

A pub

A pub (primitive unified bloc) is (circuit, observables). Observables can be a single SparsePauliOp or an array of them:

from qiskit import QuantumCircuit
from qiskit.quantum_info import SparsePauliOp

from menos.qiskit import MENOSEstimator

qc = QuantumCircuit(2)
qc.h(0)
qc.cx(0, 1)

estimator = MENOSEstimator()
job = estimator.run([(qc, [SparsePauliOp("ZZ"), SparsePauliOp("XX"), SparsePauliOp("ZI")])])
result = job.result()[0]
print(result.data.evs)

assert list(result.data.evs) == [1.0, 1.0, 0.0]
assert float(result.data.stds.max()) == 0.0

Several pubs in one run are evaluated in order, and job.result() has one PubResult per pub.

Precision

The engines are exact, so stds is always zero. A requested precision is recorded in each result's metadata and never used to add noise:

from qiskit import QuantumCircuit
from qiskit.quantum_info import SparsePauliOp

from menos.qiskit import MENOSEstimator

qc = QuantumCircuit(1)
qc.h(0)
result = MENOSEstimator(default_precision=0.05).run([(qc, SparsePauliOp("X"))]).result()[0]
print(result.metadata)
assert result.metadata["target_precision"] == 0.05
assert abs(float(result.data.evs) - 1.0) < 1e-12

Parameters must be bound first

MENOSEstimator rejects any circuit that still has parameters, even if you pass parameter_values in the pub. Bind them with assign_parameters beforehand, as the EstimatorV2 example does. The error is UnsupportedFeatureError:

MENOSEstimator does not currently support parameterized circuits (1 unbound parameter(s)); bind all parameters first (QuantumCircuit.assign_parameters).

How it runs

For each pub, MENOS converts the circuit once and then performs one planned, expectation-only run per observable element. Each run is planned independently, so different observables can be executed by different engines. Terms of one SparsePauliOp are evaluated together in a single run.

Differences from Qiskit's reference estimator

	MENOSEstimator	StatevectorEstimator
Result	exact	exact
Parameterized circuits	rejected	supported
Non-Hermitian observable	real part returned	real part returned
Scale	engine-dependent; can exceed statevector limits	statevector-limited

See also

• Expectation Values
• Estimator reference
• Supported Observables