This update makes comparing quantum results clearer and more trustworthy. WHAT'S NEW - Instant comparison summary: for every run you get the ideal outcomes, the share of shots that landed on them, the most common outcomes, the excess TVD and a plain-language verdict (close to ideal, moderately noisy, or substantially degraded). - Side-by-side ranking: runs are ordered by how far each is from its own ideal, with device and shot counts shown for context. - Different circuits, explained: when you compare different circuits, the app says so and judges each run against its own ideal. - Borderline warnings: results near a quality line show how close they are. - Uneven split detection: when the ideal outcomes are equally likely (for example, GHZ states), the app shows how uneven the observed split is. - Optional AI takeaway: one short sentence from the on-device model, verified against the calculated numbers before it appears. BETTER RELIABILITY - Comparison numbers are computed by the app itself, so they no longer depend on the AI model. - The summary shows even when the AI model is unavailable. Available on iPhone, iPad and Mac.

FROM QUANTUM EDUCATION/DEBUGGING TO QUANTUM RESEARCH EGK Quantum now introduces a dedicated Research chapter alongside Education and Debugging. NEW - Persistent Quantum Experiment Notebook - JSON and CSV research export - Ideal local statevector reference - Hardware-vs-ideal TVD and fidelity comparison - Hellinger and Jensen-Shannon diagnostics - Raw measurement-count preservation - Calibration snapshot capture - IBM calibration-aware qubit/layout ranking - Local readout-error mitigation - Pre-submission QPU workload estimate - Cross-provider IBM/OQTOPUS benchmarking - Curated educational circuit lessons - Platform-aware local simulation limits PROVIDER RELIABILITY - Updated OQTOPUS job lifecycle to its current registered-job/presigned-upload/submit model - Added OpenQASM 2 to OpenQASM 3 conversion for the supported OQTOPUS circuit subset - Updated OQTOPUS status and result handling - IBM calibration/property capture aligned to the current backend REST surface - Observed shot counts now derive from returned histograms when available RESEARCH INTEGRITY All numerical research calculations are performed deterministically in Swift. EGKAICore provides explanation and narration but is not used to calculate research metrics or calibration rankings. ROADMAP Zero-noise extrapolation is identified as the next mitigation stage and is not falsely represented as implemented in this release.

Compare Runs now includes a statistical comparison panel, showing how far apart your results are (total variation distance, fidelity, Jensen-Shannon divergence, and new outcomes) when you select two or more runs. When a circuit uses qubits that aren't directly connected on the chosen IBM device, the error message now shows how far apart they are and roughly how many SWAP operations routing would need. Under-the-hood improvements to how EGK Quantum reads IBM backend information.

This update turns EGK Quantum into a full AI-assisted quantum workbench: - Describe a circuit in your own words — the on-device model drafts the QASM - Ask AI to optimize any circuit for fewer gates before you spend real QPU time on it - Every run is now saved automatically — search and revisit your job history anytime - Compare 2-4 past results side-by-side with a visual chart and an AI-written summary All new AI features run entirely on-device — nothing leaves your phone.

Bug Fixes & Improvements Fixed circuit failures on IBM Quantum for Teleportation, Grover's Algorithm, and Quantum Fourier Transform (QFT) circuits Fixed AI explanations sometimes failing to appear after the app returned from the background during a real quantum hardware queue wait Fixed a Dark Mode issue where text was unreadable General reliability improvements to circuit submission and on-device explanation