The Ouroboros of Benchmarking: Reasoning Evaluation in an Era of Saturation

İbrahim Ethem Deveci, Duygu Ataman

The rapid rise of Large Language Models (LLMs) and Large Reasoning Models (LRMs) has been accompanied by an equally rapid increase of benchmarks used to assess them. However, due t…