Multi-Level Validation of autoq-qec: Analytical Prediction, Noise-Calibrated Simulation, and Real Hardware Execution on IBM Fez

This dataset and report present a three-level validation of the autoq-qec quantum error correction resource estimator (v3.4.3): (1) analytical prediction, (2) five noise-calibrated simulation runs, and (3) five independent executions on real quantum hardware (IBM Fez, Heron r2, 156 qubits), using a quantum Pi estimator circuit as benchmark. Across five real-hardware trials, the autoq-qec fidelity prediction agreed with measured fidelity within 0.821 +/- 0.131 percentage points, consistently inside a 1% consistency threshold. Simulation runs showed error roughly 23x-40x smaller than real hardware, quantifying the gap between simple depolarizing noise models and real hardware noise. Includes the full validation report (Markdown), the reproducible validation script, and raw JSON results for all 10 runs (5 simulation + 5 real hardware, with publicly verifiable IBM Quantum job IDs). v2: replaced simulation run 1 with a freshly-generated run backed by a verifiable raw JSON file (the original run's file had been accidentally overwritten before archiving).