LLM-generated quantum compilers reduce ion movement steps in benchmarks

Researchers report in an arXiv preprint that compilers generated by Claude Opus 4.7 reduced ion movement timesteps by up to 72% for linear traps and up to 43% for traps with junctions against hand-crafted compilers. These programs translate quantum algorithms into sequences of ion movements within a processor. The model generated and iteratively refined Python code from written specifications, using earlier compilers to develop versions for more complex architectures. Repeating the process with Claude Fable 5 reproduced the findings. The researchers say development took a few days instead of several months. For freely connected trap architectures, movement requirements varied substantially with connectivity.