Super HN

New Show
   LongCodeBench: Evaluating Coding LLMs at 1M Context Windows (arxiv.org)