EigenOrbit Library

Quantum computing, with the classical baseline in view

Where quantum methods stand today, measured rather than promised: what each method is, where it applies in industry, and how it compares with the classical method it would replace. Every page comes with eigenkit code to rerun the test on your own data, and every result names the run or the paper it comes from. EigenOrbit Benchmarks shows which of today's QPUs could run a given circuit and at what price, and EigenOrbit Router runs it on one.

Applications

Industry use cases, each with the classical method it has to beat.

Application

Molecular optimization with VQE: bond lengths of H₂ and LiH against exact chemistry

Can VQE find a bond length? We scanned H₂ and LiH on a simulator: exact from a Hartree-Fock start, trapped on LiH from random angles, slower than FCI.
Application

Quantum machine learning for wind power forecasting: a hybrid quantum network against persistence and classical models

Can quantum machine learning forecast wind power? We tested a 5-qubit hybrid network on two years of Danish grid data, 1 and 6 hours ahead. How to test yours.
Application

Quantum machine learning for fraud detection: a hybrid quantum classifier at the real fraud rate

Does quantum machine learning detect card fraud better? We tested a 5-qubit hybrid model at the real 0.17% fraud rate against classical ones. How to test yours.
Application

Quantum reservoir computing for volatility forecasting: what the evidence says

Can a quantum reservoir forecast volatility? We tested a 6-qubit reservoir against HAR on two series and checked the published S&P 500 study. How to test yours.
Application

Quantum computing for option pricing: three routes and when each is real

Can quantum computers price options? Amplitude estimation needs quadratically fewer samples but a fault-tolerant machine; tensor networks run on GPUs now.
Application

Quantum portfolio optimization: the three routes, what each has shown, and what to run instead

Does quantum portfolio optimization work yet? We ran QAOA on a 14-asset selection against exact solvers and reviewed all three routes. How to test yours.
Application

VaR and CVaR with quantum amplitude estimation: a practical guide for risk teams

Can amplitude estimation compute VaR and CVaR? We ran 7-qubit IQAE against Monte Carlo: quadratically fewer queries, not yet a shorter risk run.
Application

Breast cancer classification with a quantum SVM: a benchmark against classical models

Can a quantum SVM classify breast cancer better? We tested 4- and 8-qubit quantum kernels against tuned classical models over 10 splits. How to test yours.
Application

Quantum machine learning for defect detection: a quantum kernel SVM against classical models on steel plate faults

Can quantum machine learning classify steel defects? We tested 4- and 8-qubit quantum kernels against classical models on 1,941 plates. How to test yours.
Application

Quantum machine learning for predictive maintenance: a hybrid quantum classifier against classical models on AI4I 2020

Can quantum machine learning predict machine failures? We tested a 5-qubit hybrid classifier on AI4I 2020 at the real failure rate. How to test yours.

Comparisons and guides

Which method to pick, and how to run a fair comparison yourself.

Guide

Quantum computing for chemistry: what has run, what beats it today, and what it would take

Quantum computing for chemistry, by the evidence: what has run on hardware, how it compares with classical methods, and what FeMoco and drug design need.
Guide

Quantum simulation of materials and quantum matter: what quantum computers have done, and what classical computers matched

Quantum simulation of materials and many-body physics: every beyond-classical claim, which ones classical methods matched, and which still stand.
Comparison

eigenkit, marqov, quantum-executor and SuperQuantX: four ways to run one circuit on many backends

Four open-source Python libraries that run one quantum program on many backends: what eigenkit, marqov, quantum-executor and SuperQuantX each do, and which to pick.
Comparison

Does a quantum layer help? HQNN vs a classical network on the same data

Does a quantum layer help a neural network? We tested a 5-qubit hybrid against a classical twin of the same size on two datasets. How to test yours.
Comparison

Monte Carlo or quantum amplitude estimation: when the quadratic speed-up is real

Does amplitude estimation beat Monte Carlo? We measured IQAE's quadratic query advantage on a simulator, and what still stands between it and less time.
Comparison

QAOA vs MILP: when an exact classical solver beats a quantum heuristic

QAOA vs MILP: a mixed-integer solver proves its optimum, QAOA samples candidates. We ran both on one 14-asset problem: what it shows and what could flip it.
Comparison

Qiskit vs Cirq vs PennyLane: which quantum SDK to choose

Qiskit, Cirq or PennyLane? What each quantum SDK is built for, which hardware it reaches, how the same circuit looks in all three, and how to choose.
Comparison

Quantum kernel vs RBF kernel: how to run a fair QSVM benchmark

QSVM vs SVM: we tested 4- and 8-qubit quantum kernels against tuned RBF and linear SVMs on breast cancer data. How to run a fair benchmark, with code.
Comparison

Quantum annealing vs QAOA: two ways to solve the same QUBO, and how to choose

Quantum annealing vs QAOA: both minimise a QUBO, one on analog annealers, one with gates. How they differ, how to choose, and a worked example on one problem.
Comparison

Quantum annealing vs simulated annealing vs simulated bifurcation: what the benchmarks show

Quantum annealing, simulated annealing, simulated bifurcation, digital annealers and CIMs take the same QUBO. What benchmarks found, and how to test yours.
Comparison

PennyLane vs Qiskit and other quantum machine learning frameworks, and where eigenkit sits

PennyLane vs Qiskit, CUDA-Q, Braket, TensorCircuit and Qulacs for quantum machine learning: licence, gradients, hardware, PyTorch and scikit-learn support.
Comparison

Superconducting vs trapped-ion vs neutral-atom qubits: how the types of quantum computer compare

Superconducting vs trapped-ion vs neutral-atom qubits, plus photonic and analog machines: gate fidelity, speed, connectivity and size of named QPUs today.
Guide

How much does quantum computing cost? QPU prices, billing models and worked examples

How much does quantum computing cost? QPU prices from IBM, AWS Braket, Azure Quantum and IQM as of November 2026, the billing models and worked examples.
Guide

Is quantum machine learning useful? When it makes sense, and when it does not

Is quantum machine learning useful? For most problems on classical data today, no. Where it makes sense, and how a benchmark on your own data decides it.
Guide

Quantum computing use cases: where quantum computers will help first, ranked by the evidence

Quantum computing use cases ranked by evidence: chemistry, materials and physics simulation first, then codebreaking, with optimisation and ML far behind.
Guide

Quantum computing for logistics: routing and scheduling, and what the trials show

Routing, scheduling and loading can be written for a quantum computer. What the published trials ran, how fast the qubit count grows, and what to do today.

Concepts

The ideas under quantum computing, with the circuits that show them.

Concept

Ansatz: the trial circuit behind every variational quantum algorithm

An ansatz is the parametrised circuit a variational quantum algorithm searches over. The main types, the trade-off between them, and how to choose one.
Concept

Barren plateaus: why quantum neural networks stop training

A barren plateau is a training landscape whose gradients vanish exponentially as qubits are added. What causes it, how to avoid it, and the catch in doing so.
Concept

Quantum advantage: what it means, every major claim, and what has held up

Has quantum advantage been achieved? Yes for tasks built to test it, not yet for useful ones. Every major claim since Sycamore, and what held up.
Concept

Quantum error correction: logical qubits, the threshold, and how far hardware has come

Quantum error correction turns noisy qubits into logical qubits. What below threshold means, which machines have crossed it, and what is still missing.
Concept

Quantum neural networks (QNN) explained: QNN vs hybrid QNN (HQNN), and how to test whether the quantum layer helps

A quantum neural network is a trainable quantum circuit used as a model. How a QNN differs from a hybrid QNN (HQNN), and how to test whether the quantum layer helps.
Concept

Quantum superposition: what a qubit holds before you measure it

What quantum superposition is: a qubit holds an amplitude on 0 and on 1, measurement returns one of them, and interference is what algorithms use. With eigenkit code.

Methods

How to apply a quantum method to a problem. The eigenkit docs have the API.

Method

Grover's algorithm explained: fewer queries, not always less time

Grover's algorithm finds one marked item among N in about (π/4)√N oracle calls, not N/2. What that means in wall-clock time, and where the idea is used today.
Method

HHL algorithm: does your linear-system problem qualify?

The HHL algorithm prepares a quantum state proportional to the solution of Ax = b, and helps only if four conditions hold. A checklist with resource estimates.
Method

QAOA in practice: how to write a real problem for it, what to compare it with and how to read the results

QAOA on a real problem: constraints become penalties, exact solvers set the bar, and sampled shots are the result. Tested on a 14-asset selection.
Method

Quantum circuit Born machines: generative models that sample by measurement

A quantum circuit Born machine is a generative model whose measurement outcomes are its samples. How it trains, how it compares with QGANs and MPS, with code.
Method

The quantum Fourier transform explained: a cheap circuit you cannot read out

The quantum Fourier transform takes about n²/2 gates on n qubits, where an FFT takes n × 2ⁿ steps. What that buys, why it is not a faster FFT, and where it is used.
Method

QUBO and Ising: how to write an optimisation problem for quantum annealing and QAOA, and how to choose a solver

Quantum annealing minimises a QUBO: a quadratic cost of binary variables. How to write constraints as penalties, check it in eigenkit and pick a solver.
Method

Simulated bifurcation: a classical, quantum-inspired solver for Ising and QUBO problems

Simulated bifurcation is a classical, quantum-inspired heuristic for Ising and QUBO problems. How it works, what benchmarks show, and Toshiba's trading tests.
Method

Tensor networks: matrix product states, tensor trains and what they compute

Tensor networks explained: how matrix product states and tensor trains compress huge arrays, why they simulated a 127-qubit IBM run, and where they fail.
Method

The variational quantum eigensolver (VQE) explained: the loop, its costs and its limits

VQE tunes a quantum circuit to lower a measured energy. How the loop works, what it costs in measurements, where it fails, and what a run of ours shows.