Mohammad Arif Rasyidi

Benchmarking API Drift in LLM-Generated Quantum Code Across Successive SDK Versions

Mohammad Arif Rasyidi, Syahirul Faiz

Large language models can generate plausible quantum code, but it is unclear whether they can reliably target the specific software development kit (SDK) version requested by the user. We study this problem as API drift and introduce quantum-api-drift, a benchmark for measuring v…Read more