Multi-LCB extends LiveCodeBench to multiple programming languages

arXiv·medium signal

Multi-LCB broadens LiveCodeBench, a widely adopted code-generation benchmark, beyond its original scope to evaluate LLMs across multiple programming languages. Since LiveCodeBench is a common reference for ranking coding agents and models, multi-language coverage exposes how much current results are skewed toward Python and a few popular languages. For builders choosing a coding agent, it provides a more honest cross-language picture of capability.