The Cirq developer glossary
Twenty Cirq terms that map Google Quantum AI’s Python stack, from Moments and qubit types to the cloud Engine. Every definition here is checked against the Cirq API as Google ships it today rather than legacy tutorials.
The essential vocabulary for Google Quantum AI’s Python quantum framework
Cirq is Google Quantum AI’s open-source Python framework for designing, simulating, and running quantum circuits, and the native SDK for Google’s superconducting processors, from the 53-qubit Sycamore chip of 2019 to the 105-qubit Willow announced in December 2024. Cirq makes the structure of a quantum circuit unusually explicit: every program is a sequence of Moments, every operation is bound to a specific qubit, and the framework is honest about hardware constraints from the start. These 20 Cirq terms are the vocabulary you need to read Cirq code, choose the right qubit type and simulator for the job, and write circuits that will actually run on Google Quantum AI hardware via the cloud Engine. Sister glossaries: Top 20 Qiskit Terms and Top 20 Q# Terms.
Cirq
Cirq is Google Quantum AI’s open-source Python framework for designing, simulating, and running quantum circuits. It is built around an explicit, low-level circuit model and is the native SDK for Google’s superconducting processors, including the Sycamore and Willow chips, accessed through Google Quantum Engine.
Circuit
The central object in Cirq. A cirq.Circuit is an ordered collection of Moments, each of which contains operations that act on disjoint qubits. Circuits in Cirq are intentionally explicit about parallelism, what executes together belongs in the same Moment.
Moment
A single time-slice of a circuit. Every operation inside a Moment acts on different qubits and is taken to execute in parallel. Cirq’s Moment-based model gives you precise control over circuit timing and depth, which matters on noisy hardware.
Operation
A Gate applied to specific qubits, e.g. cirq.X(q0) or cirq.CNOT(q0, q1). Operations are the elements that populate a Moment.
Gate
A reusable, qubit-agnostic quantum operation such as cirq.H, cirq.X**0.5 or cirq.CZ. Cirq exposes a rich library of standard gates and supports raised-power syntax such as X**t for fractional rotations natively.
Qubit
An abstract qubit identifier. Cirq does not assume a single qubit type; you choose one that matches your problem: NamedQubit for symbolic work, LineQubit for 1D arrangements, and GridQubit for 2D layouts that match real hardware.
GridQubit
A qubit identified by its (row, column) position on a 2D grid, designed to match the topology of superconducting hardware such as Sycamore. Using GridQubits lets the simulator and compiler reason about device-realistic connectivity.
LineQubit
A qubit identified by a single integer position on a 1D line. LineQubits are convenient for prototyping, teaching, and any algorithm whose connectivity is one-dimensional.
Measure
A measurement operation, added to a circuit with cirq.measure(q, key='m'). Each measurement is associated with a string key, and that key is how you retrieve the corresponding bitstring from a Result after sampling.
Simulator
Cirq’s default statevector simulator (cirq.Simulator). It supports both exact final-state computation and sampled measurements, and is the right choice for noiseless simulation of small-to-medium circuits.
DensityMatrixSimulator
A simulator (cirq.DensityMatrixSimulator) that evolves the system’s density matrix instead of a pure statevector, allowing you to model noise channels, mixed states, and decoherence. Essential for studying how realistic noise will affect an algorithm.
Sampler
An abstraction for anything that can run circuits and return measurement results, cirq.Simulator is one example, the cloud Engine is another. Writing code against the Sampler interface lets you swap simulator and hardware without changing your algorithm code.
Result
The object returned by a sampler run. A Result contains the measurement outcomes for every measurement key, exposed as histograms, raw bitstrings, or pandas DataFrames depending on what you ask for.
PauliString / PauliSum
Cirq’s representation for Pauli operators and Hamiltonians. A PauliString is a tensor product of single-qubit Paulis with a coefficient; a PauliSum is a weighted sum of these. Both are the input format for expectation-value calculations and many algorithm libraries.
PhasedXZGate
A hardware-native parameterised single-qubit gate (cirq.PhasedXZGate) that encapsulates the most general single-qubit unitary in a form Google’s hardware can execute directly. Important when targeting Sycamore-class devices, where decomposing into hardware-native gates produces shallower, lower-noise circuits.
Decomposition
Expressing a gate or operation in terms of other, lower-level gates, for example, decomposing a Toffoli into one- and two-qubit gates from a hardware basis set. Decomposition is how Cirq compiles arbitrary unitaries down to what a target device can run.
Transformer
Cirq’s circuit-rewriting and optimization API. Transformers (cirq.transformers) include passes for merging gates, dropping negligible operations, aligning operations into Moments, and applying device-specific optimisations. They are the modern replacement for the older optimizer abstraction.
Device
An object that describes a hardware target’s constraints, its supported gate set, qubit grid, allowed two-qubit interactions, and timing. Validating a circuit against a Device tells you whether it will run on the actual hardware before you submit it.
Engine
cirq_google.Engine is the client interface for Google Quantum AI’s cloud service, used to submit jobs to Google’s superconducting processors such as Willow and receive results. Engine handles authentication, project scoping, processor selection, and job lifecycle management.
Cirq-FT
cirq-ft was Cirq’s fault-tolerant extension, providing primitives, Bloqs, T-gate accounting, and resource estimation for algorithms aimed at future error-corrected hardware. Google Quantum AI has since developed Qualtran as the actively maintained successor for expressing and resource-estimating large-scale fault-tolerant algorithms, so new work should build on Qualtran rather than the legacy cirq-ft module.
How these Cirq terms fit together
The twenty Cirq terms above build on one deliberate design choice, that a circuit is an explicit sequence of Moments. You place Operations, each a Gate bound to a Qubit, into Moments that execute in parallel, choose GridQubit or LineQubit identifiers to match your hardware layout, and then hand the finished Circuit to a Simulator or to the cloud Engine through the Sampler interface. Once that spine is clear, the rest of the vocabulary, from DensityMatrixSimulator noise studies to Transformer optimisation passes, fits naturally around it.
These Cirq terms sit alongside the vocabularies of the other major SDKs, and the underlying concepts translate well between them. If you also target IBM hardware, read our Qiskit terms glossary, and for Microsoft tooling see the Q# terms glossary. This page is one of three SDK companions inside the Quantum Books hub, where the full reading path runs from first principles to production code. For a wider view of where Cirq sits among the options, see our comparison of quantum programming languages.
Common questions about Cirq terms
Which Cirq terms should a beginner learn first
Begin with Circuit, Moment, Operation, and Gate, because those four define how every Cirq program is structured. Add Qubit, Simulator, and Result next, and you will have enough vocabulary to write a small circuit and read its measurement outcomes.
Why does Cirq make you choose a qubit type
Cirq keeps qubit identity explicit so your circuit can reason about real hardware topology from the start. LineQubit suits one-dimensional prototypes, while GridQubit encodes the row and column layout of superconducting chips such as Willow, which helps the compiler respect device connectivity.
Is cirq-ft still the tool for fault-tolerant work
The legacy cirq-ft module still exists, but Google Quantum AI now develops Qualtran as its successor for fault-tolerant algorithm design and resource estimation. New projects that need Bloqs, T-gate accounting, and large-scale resource counts should build on Qualtran rather than cirq-ft.
How current is this Cirq terms glossary
Every definition reflects the current Cirq framework and its cloud Engine as of July 2026, including the shift toward Willow-class hardware. We revisit these Cirq terms each time Google cuts a significant Cirq release so the vocabulary stays aligned with the API you actually import.
External reference: for the canonical specification, see the official Google Cirq documentation reference for cirq terms.
See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.
