IBM's open source Qiskit Paulice add on for IBM's quantum development framework scatters lightweight error checks across a quantum circuit's execution timeline and discards failed runs before analysis — a postselection approach that trades the number of circuit executions (shot…
Today's quantum computers waste most of what they compute. The chips misfire often enough that researchers running experiments routinely throw away the bulk of their results, since the runs fail quality checks before they ever reach analysis. That discarded work is the real bottleneck for any team operating on current hardware, and it is the problem IBM's new open-source package, Qiskit Paulice, is built to attack.
Qiskit Paulice is a small software add-on for Qiskit, IBM's open-source quantum development framework. Instead of waiting for the dedicated error-correction hardware IBM expects to arrive around 2029, the package scatters lightweight error checks across a quantum circuit's execution timeline. Each check inspects the program's state at a specific moment. If a run has gone wrong, the tool flags it for postselection, the practice of discarding failed runs and keeping only the clean ones. The effect is that practitioners recover more usable runs from the same noisy hardware, paying the cost in discarded shots rather than in extra computation per shot.
The underlying technique is documented in a companion preprint, "Low-overhead error detection with spacetime codes" (arXiv 2504.15725). The package was co-developed by IBM researchers Simon Martiel and Ali Javadi-Abhari, and it is available as a public GitHub repository under Qiskit/qiskit-paulice, with official documentation and a walkthrough guide on low-overhead error detection with spacetime codes.
What Paulice is not is worth marking as clearly as what it is. It is not full quantum error correction; that still requires redundant logical qubits and dedicated hardware IBM has not yet deployed. It is also not a replacement for two popular error-mitigation techniques already in Qiskit: zero-noise extrapolation (ZNE) and probabilistic error cancellation (PEC), both of which demand exponentially more sampling time. Paulice's pitch is that spacetime checks add only modest gate and qubit overhead, so the cost is paid in discarded runs rather than in additional compute per run.
That cost framing is what makes the trade-off legible. Postselection means a researcher running a 1,000-shot experiment might keep only a fraction of the data if error rates run high. The package is useful precisely when the alternative is throwing the entire experiment away anyway, but it does not pretend to eliminate noise. It shifts the budget from raw compute to useful output.
The scope is also narrow by design. Paulice is built for Clifford circuits, a widely used family of quantum programs that includes many benchmarking and error-detection routines, but not all algorithms fit the category. Practitioners building non-Clifford workloads will need to keep waiting, whether for later add-ons, for the 2029 fault-tolerant target, or for hardware-native error correction.
The release was covered by Quantum Computing Report as an open-source tooling story rather than a hardware milestone, which is closer to what is actually in the repository. The watch item now is whether teams running on IBM's cloud quantum hardware and other current-generation platforms adopt the add-on into routine workflows, and whether the underlying spacetime-check technique extends beyond Clifford circuits in future releases.