LiveOIBench: Can Large Language Models Outperform Human Contestants in Informatics Olympiads?
LiveOIBench introduces a rigorous evaluation framework for Large Language Models in informatics Olympiads, revealing that GPT-5's performance remains slightl...