QuaMap Publishes 210000 IBM Quantum Circuits for Compiler Benchmarking

The open QuaMap dataset links 75 algorithms to more than 210000 transpiled circuits across IBM backends, giving ML compiler teams a shared layout ranking benchmark.

Key points

Secbrain released QuaMap on GitHub as an open dataset and benchmark suite for quantum circuit mapping.
The README states it connects 75 quantum algorithms to more than 210000 hardware aware transpiled circuits across IBM Quantum backends, with metrics such as depth, CX counts, gate counts, layout metadata, and mapped QASM.
Quantum software teams increasingly use learning based compilers, but progress stalls without shared, hardware faithful datasets.

By Classy AI News

Published October 14, 2026

What changed

Secbrain released QuaMap on GitHub as an open dataset and benchmark suite for quantum circuit mapping. The README states it connects 75 quantum algorithms to more than 210000 hardware aware transpiled circuits across IBM Quantum backends, with metrics such as depth, CX counts, gate counts, layout metadata, and mapped QASM. A Zenodo DOI badge points to archived release 10.5281/zenodo.18154917. Tasks include layout ranking, transpilation metric prediction, and cross device transferability analysis.

The repository layout separates origin circuits, transpiled circuits grouped by backend, and JSON metric files per layout. That structure matters for teams training graph models because it preserves the pairing between logical algorithms and hardware specific mappings rather than dumping undifferentiated QASM files.

Why it matters

Quantum software teams increasingly use learning based compilers, but progress stalls without shared, hardware faithful datasets. QuaMap encodes real backend topologies and exhaustive small qubit layout enumeration, which lets R&D groups compare models on prediction tasks that mirror production transpiler pain. For capital allocators, the artifact is a sign that compiler ML is moving from toy graphs to IBM scale circuit corpora.

If your roadmap assumes proprietary logs alone will defend margin, public corpora like QuaMap compress the time competitors need to reproduce baseline layout rankers. The decision is whether to contribute upstream metrics or keep internal telemetry as the differentiator.

Who is affected

Quantum software engineers, Qiskit contributors, researchers benchmarking variational and fault tolerant era compilers, and cloud quantum product managers evaluating customer workload portability.

What to do next

Clone the repository, inspect the metrics JSON schema for one algorithm family, and attempt a baseline layout ranking model before committing roadmap time to custom internal logs only benchmarks.

What to watch

New backend entries as IBM refreshes processors, plus peer reviewed citations that report transferability gaps between Heron era devices.

Sources

Primary — Secbrain, QuaMap README on GitHub (2026). Dataset scope, structure, and tasks.
Primary — Zenodo, QuaMap archive DOI 10.5281/zenodo.18154917 (2026). Citable release artifact.