Logical qubit
Also known as: logical qubits, error-corrected qubit, fault-tolerant qubit, encoded qubit
A logical qubit is an error-protected qubit created by combining many physical qubits with quantum error correction.
Physical qubits are noisy, but a well-built logical qubit can be far more reliable. While headlines often cite physical qubit counts, logical qubits are the number that actually determines what a quantum computer can do, and today's machines have very few.
Logical qubit, explained at three levels
At home
A physical qubit is like a single unreliable worker who sometimes makes mistakes. A logical qubit is like a team that checks each other's work, so the final result is reliable. Building one logical qubit takes many physical qubits.
At work
When evaluating quantum computing news, logical qubits are the number to watch. Useful applications like breaking encryption or simulating complex molecules are estimated to need hundreds to thousands of logical qubits, far more than exist today.
At home
The number of physical qubits per logical qubit depends on how error-prone the physical qubits are and which error correction code is used. Better physical qubits mean fewer are needed for each logical one. That's why both quality and quantity matter.
At work
Logical qubits are characterized by logical error rate per operation or per round of error correction. Claims of 'logical qubits' vary in rigour: some demonstrations detect errors without correcting them, while others run repeated real-time correction. Ask how errors were handled and whether performance improves as code size increases.
At home
Logical operations must also be fault tolerant, meaning a single physical error can't spread and corrupt the logical result. In my view, that's the subtle part: it's not enough to store information safely, you have to compute on it without errors cascading.
At work
Resource estimates for useful algorithms often specify logical qubits, logical error rates, and T-gate counts. For example, a 2025 estimate suggested breaking RSA-2048 could take under a million noisy physical qubits over about a week, far fewer than earlier estimates, but still well beyond current hardware. Roadmaps from IBM, Google, Quantinuum, QuEra, and others target growing numbers of logical qubits through the late 2020s.
Real-world examples
At home
Carlos saw an announcement about a quantum computer with over 1,000 qubits and assumed it could break encryption. He learned that those were physical qubits, and that breaking encryption would need many error-corrected logical qubits, which didn't yet exist.
At work
Lena is a technology strategist at a pharmaceutical company. When vendors pitched quantum chemistry pilots, she asked how many logical qubits their simulations required and when hardware would deliver them, which helped set realistic expectations with leadership.
How it compares
| feature | Logical qubit | Physical qubit |
|---|---|---|
| What it is | Error-corrected qubit built from many physical qubits | A single hardware qubit |
| Error rate | Much lower, if correction works well | Higher, limited by hardware |
| Count in today's machines | Few, in experimental settings | Dozens to over a thousand |
| Needed for useful tasks | Hundreds to thousands | Potentially millions |
| What headlines usually cite | Increasingly, in recent years | Most often |
Physical qubits are the raw material; logical qubits are the reliable building blocks that matter for useful work.
For analysts
When a quantum press release cites a big qubit count, I'd immediately ask whether it's physical or logical. That single question filters out most of the hype in quantum computing news.
History & origin
The idea of logical qubits came with the first quantum error correction codes in the mid-1990s. For years, experiments demonstrated pieces of error correction without producing logical qubits that outperformed their physical components. In December 2023, a team from Harvard, MIT, and QuEra reported a neutral-atom processor running algorithms on up to 48 logical qubits, though with limited error correction. In 2024, Microsoft and Quantinuum reported logical qubits with error rates far below physical ones, and Google's Willow chip showed logical errors shrinking as code size grew. Vendor roadmaps, such as IBM's plan for a fault-tolerant system around 2029, now centre on logical qubit targets. In my opinion, the shift to talking about logical qubits is the clearest sign the field is maturing.