Search papers, labs, and topics across Lattice.
QuanBench+ is introduced as a unified benchmark for evaluating LLMs on quantum code generation across Qiskit, PennyLane, and Cirq, featuring 42 aligned tasks. The benchmark uses executable functional tests and KL-divergence for probabilistic outputs to evaluate models, reporting Pass@1 and Pass@5 metrics. Results indicate that while LLMs achieve reasonable performance, especially with feedback-based repair, significant framework-specific knowledge is still required for reliable multi-framework quantum code generation.
LLMs still struggle to generalize quantum code generation across frameworks like Qiskit, PennyLane, and Cirq, even with feedback-based repair pushing performance to 83.3%, 76.2%, and 66.7% respectively.
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.
