Qubit Error Budgets
An error budget accounts for how each physical mechanism contributes to a gate's total error, guiding where engineering effort will pay off.
Why Budget Errors
A measured gate infidelity is a single number, but it is the sum of many contributions: coherence-limited decay, leakage, coherent control error, crosstalk, and readout error. Improving the gate requires knowing which contribution dominates, because effort spent on a small term is wasted. An error budget decomposes the total error into named physical sources with estimated sizes, turning a scalar metric into an actionable diagnosis.
The Coherence Floor
For a gate of duration t_g on qubits with relaxation time T1 and dephasing time T2, there is an unavoidable error from decoherence during the gate. A useful rule of thumb for the coherence-limited error of a two-qubit gate is roughly proportional to t_g times the sum of the relevant decay rates. This sets a floor: even a perfectly calibrated gate cannot do better than its coherence allows, so faster gates or longer coherence are the only ways to lower the floor.
- Incoherent error from T1 and T2 during the gate, the coherence floor.
- Coherent error from miscalibration: over-rotation, axis tilt, phase error.
- Leakage out of the computational subspace.
- Crosstalk and spectator-induced error from neighboring qubits.
- Readout and state-preparation error where relevant.
How the Terms Are Found
Coherence times are measured directly and converted to an expected floor. Coherent errors are extracted by gate-set tomography or by purity benchmarking, which separates coherent from incoherent contributions. Leakage is measured with dedicated leakage benchmarking. Crosstalk shows up in simultaneous benchmarking. When these estimated terms add up to the measured total error, the budget is closed and trustworthy.
Using the Budget
A closed budget points to the next investment. If the coherence floor dominates, better materials or faster gates help. If coherent error dominates, recalibration or pulse shaping helps. If leakage dominates, pulse trajectories and reset schemes help. Error budgets are how hardware teams decide priorities, and they connect every mechanism described in this library to the bottom-line gate fidelity that ultimately determines whether error correction can succeed.