QuanBench+: A Unified Multi-Framework Benchmark for LLM-Based Quantum Code Generation

TL;DR

QuanBench+ introduces a comprehensive benchmark for evaluating large language models on quantum code generation across multiple frameworks, highlighting progress and ongoing challenges.

Contribution

It provides a unified multi-framework benchmark with aligned tasks and evaluates models with new metrics, including feedback-based repair, revealing framework-specific strengths and weaknesses.

Findings

Strongest one-shot scores: Qiskit 59.5%, Cirq 54.8%, PennyLane 42.9%.

Feedback-based repair improves scores significantly.

Quantum code generation remains dependent on framework-specific knowledge.

Abstract

Large Language Models (LLMs) are increasingly used for code generation, yet quantum code generation is still evaluated mostly within single frameworks, making it difficult to separate quantum reasoning from framework familiarity. We introduce QuanBench+, a unified benchmark spanning Qiskit, PennyLane, and Cirq, with 42 aligned tasks covering quantum algorithms, gate decomposition, and state preparation. We evaluate models with executable functional tests, report Pass@1 and Pass@5, and use KL-divergence-based acceptance for probabilistic outputs. We additionally study Pass@1 after feedback-based repair, where a model may revise code after a runtime error or wrong answer. Across frameworks, the strongest one-shot scores reach 59.5% in Qiskit, 54.8% in Cirq, and 42.9% in PennyLane; with feedback-based repair, the best scores rise to 83.3%, 76.2%, and 66.7%, respectively. These results show clear progress, but also that reliable multi-framework quantum code generation remains unsolved and still depends strongly on framework-specific knowledge.