New benchmark reveals AI models default to Python even when other languages fit better

Original: LangChoiceBench: Measuring and Explaining Programming-Language Choice in LLMs

Source: arxiv.org ↗