Build your first quantum circuit this week
You do not need a physics degree or a quantum computer to start programming one. In 2026 the major toolkits are free and open source, several companies let you run small circuits on real hardware at no cost, and a browser is enough to begin. The skill that grows from here is simple to name: describe a problem as a circuit, run it, read noisy results honestly, and improve it. That habit points toward 2040, when error-corrected machines may need people who think in both classical and quantum code. Today's devices are still small and noisy, so we teach this as exploration and craft, not as a promise of instant speedups. This page is the B in QUBIT Quest: Building circuits.
What you'll be able to do
Build and draw a circuit
Create a Bell-state circuit in code, draw it as a diagram and explain what each gate does to the qubits.
Run it and read the results
Simulate a circuit thousands of times, plot the counts and explain why the answers are a spread of outcomes, not one number.
Try real hardware
Send a tiny job to a real quantum processor and compare its noisy results with a perfect simulator.
Choose a toolkit sensibly
Compare Qiskit, Cirq, Q#, PennyLane, CUDA-Q, pytket and others by language, goal and cost, then pick one for your next project.
Work without installing anything
Use a browser composer, a hosted notebook or Google Colab to code quantum programs on a school Chromebook.
Connect circuits to AI
Wrap a small circuit in PennyLane and train it like a model, as a first step toward quantum machine learning (see our algorithms page).
Compare the free toolkits
All of the toolkits below are free to download. The "first program" column is the exercise we suggest for your first hour. Costs for cloud hardware are noted honestly in the next section. Facts were checked against each provider's official pages in October 2026, so confirm current limits before you plan a session.
| Platform | Language | Best for | Free learning | First program |
|---|
| IBM Qiskit + IBM Quantum Platform | Python | The most widely used toolkit; real IBM hardware on the free plan | IBM Quantum Learning courses, Qiskit docs and tutorials | Bell state: H then CNOT, run on the simulator, then on a real backend |
| Google Cirq + Quantum AI | Python | Hardware-aware circuits and research-style experiments | Cirq basics and tutorials; Colab notebooks | Build a 2-qubit circuit, simulate and print the measurement histogram |
| Microsoft QDK (Q#) + Azure Quantum | Q#, plus Python, Qiskit, Cirq and OpenQASM support | A purpose-built quantum language; resource estimation | Microsoft Learn docs and quickstarts | Q# "Hello Quantum": allocate a qubit, apply H, measure, repeat |
| Amazon Braket | Python | One cloud for several hardware vendors (paid use) | AWS docs and the amazon-braket-examples notebooks | Run a Bell circuit on the free local simulator in the SDK |
| Xanadu PennyLane | Python | Quantum machine learning and hybrid quantum-classical code | PennyLane Codebook, demos and docs | Make a one-qubit variational circuit and tune a rotation angle by gradient descent |
| NVIDIA CUDA-Q | Python or C++ | Fast GPU simulation and hybrid HPC workflows | CUDA-Q Academic notebooks | Write a kernel that prepares a GHZ state and sample it |
| Quantinuum pytket | Python | Compiling and optimizing circuits for many different devices | pytket docs and user guide | Build a Bell circuit, then compile it for a simulated device and compare gate counts |
| Classiq | Python (high-level model) | Describing what you want and letting software design the circuit | Classiq docs and open GitHub algorithm library | Run one notebook from the classiq-library tutorials |
Your free learning path
Pick one toolkit first and go deep, then sample a second. Every step below is free. Hours are our own estimates. Where a certificate is not confirmed on the official page, we say so.
Step 1 · IBM Qiskit and the IBM Quantum Platform
Provider: IBM. You earn: working Qiskit skills. The IBM Quantum Learning course Basics of quantum information is the first in a four-course series, and IBM offers an optional badge exam at IBM Training; the exam price is not confirmed on the course pages, so check before you book it. Free plan: the Open Plan lists up to 10 minutes of real quantum computer (QPU) time. IBM's docs describe this as per 28-day rolling window; the pricing page words it as per month. Active users may be eligible to request more time. Cost: Free course · About 10 hours (estimate) · Level: Beginner
Open the courseSee the free plan
Step 2 · Google Cirq tutorials
Provider: Google Quantum AI. You earn: Cirq skills (no certificate). Start with Cirq basics, which builds and simulates your first circuit. Cirq is free and open source. Note that running on Google's own processors is restricted to an approved group, so treat Cirq as a simulator and learning tool for now. Cost: Free course · About 4 hours (estimate) · Level: Intermediate (assumes some quantum background)
Cirq basicsMore tutorials
Step 3 · Microsoft Quantum Development Kit (Q#)
Provider: Microsoft. You earn: Q# skills (certificate not confirmed). The QDK is free and open source, and you do not need an Azure account to use it. Follow the Q# quickstart, then the Grover search tutorial. Azure Quantum itself needs an Azure account and workspace for real hardware, and we could not confirm a current free-credits offer, so check the Azure Quantum pricing page before relying on one. Cost: Free course (QDK and docs) · About 5 hours (estimate) · Level: Intermediate
Q# quickstartAzure Quantum pricing
Step 4 · Xanadu PennyLane Codebook and demos
Provider: Xanadu. You earn: a strong foundation in quantum computing and QML through 15 listed Codebook modules, from single-qubit gates to Shor's algorithm and Hamiltonian simulation (certificate not confirmed). PennyLane runs on many simulators and hardware backends. Cost: Free course · About 20 hours (estimate) · Level: Beginner to Intermediate
Open the CodebookBrowse demos
Step 5 · NVIDIA CUDA-Q Academic
Provider: NVIDIA. You earn: hands-on skill in hybrid quantum-classical code (certificate not confirmed). CUDA-Q Academic is an open curriculum of Jupyter notebooks with a Quick Start to Quantum for beginners and advanced paths up to error correction. Notebooks can run locally or on qBraid, CoCalc or Google Colab. Cost: Free course · About 8 hours (estimate) · Level: Beginner to Advanced
Open CUDA-Q AcademicAbout CUDA-Q
Step 6 · Quantinuum pytket
Provider: Quantinuum. You earn: skill at compiling the same circuit for different devices (no certificate). pytket installs from PyPI with pip, and the docs include a user guide and getting started page. Access to Quantinuum hardware is a commercial offering; we could not confirm a free hardware tier, so use the open-source toolkit with simulators. Cost: Free course (open-source docs) · About 4 hours (estimate) · Level: Intermediate
pytket user guideGetting started
Step 7 · Classiq (optional)
Provider: Classiq. You earn: a taste of high-level quantum design (certificate not confirmed). The open classiq-library on GitHub and the docs are free to read. We could not verify a free tier for the Classiq platform itself: its academia page lists education programs but states no price, so ask Classiq or your school before counting on free access. Cost: Free course (docs and library only; platform pricing not confirmed) · About 3 hours (estimate) · Level: Intermediate
Classiq docsOpen the library
Later, when sponsored · Amazon Braket
Provider: Amazon Web Services. Braket is a paid cloud service for running circuits on several vendors' hardware. AWS lists one hour of included on-demand simulation per month for the first twelve months of using Braket, and a free local simulator in the Braket SDK. Hardware tasks are billed per use, so check the official pricing page for real prices before any session. For learners, the free local simulator and the example notebooks are the right place to begin. Cost: Free local simulator; cloud use is paid · About 4 hours (estimate) · Level: Intermediate
Braket pricing (paid)Example notebooks
Hands-on: install-free labs
These run in a browser, so they work on a school laptop. Most need a free account.
IBM Quantum Composer
Drag gates onto qubit lines, watch probabilities and the Q-sphere update, then switch to the code editor, which supports OpenQASM 2 and 3. Sign in to save your work. Try it first: make a Bell state and predict the outcomes before you press run.
Open Composer
qBraid Lab (free plan)
A web-based IDE for quantum computing with ready-made environments for machine learning, chemistry and error correction. qBraid documents a free plan with a monthly CPU-hour allowance; its pricing page and docs list slightly different limits, so read the current numbers when you sign up. Free here means free of qBraid credits: jobs on some partner devices bill to a separate account.
Check free limitsqBraid
Google Colab notebooks
Colab gives you a free hosted notebook in the browser, with nothing to install except the toolkit with one pip command. Google publishes a Cirq getting-started notebook that opens in Colab, and CUDA-Q Academic lists Colab as a way to run its lessons.
Cirq notebook in ColabGoogle Colab
PennyLane demos and the QDK playground
PennyLane demonstrations are runnable tutorials, from a first variational circuit to quantum chemistry. Microsoft's QDK also links a browser playground that opens in vscode.dev. Both let you read, change and rerun real code in minutes.
PennyLane demosMicrosoft Quantum
For Curiosity Coaches
Coaches do not need to be quantum experts. Your job is to keep learners curious, help them run the code and ask good questions.
Discussion prompts
- Why does the same circuit give different counts each run? What does that tell us about measurement?
- How is a quantum program like a recipe, and how is it not?
- The real machine gave "wrong" answers the simulator did not. What might cause that?
- If you could choose only one toolkit for a school club, which would you pick and why?
4-week plan (about 1 hour per week)
- Week 1: Composer. Build and explain a Bell state without code.
- Week 2: Recreate it in Qiskit or PennyLane in a notebook and plot the counts.
- Week 3: Run on a real device if the learner has a free IBM account, then compare with the simulator.
- Week 4: Try a second toolkit and write three sentences on which felt better.
Age and level fit: Composer and the Week 1 activity suit curious teens (about 13+) with no coding. Python notebooks suit learners comfortable with basic Python, such as high school and above. Cirq, pytket and Q# are better for intermediate learners and career changers. Always check each provider's age rules before a learner makes an account.
Free resource library
CUDA-Q, pytket and Classiq
Related on this site: Quantum Algorithms & Quantum Machine Learning, the I in QUBIT Quest.