It has some search-like characteristics, but its primary purpose is to benchmark and compare large language models (LLMs) on real-world tasks.

Is it like an search engine?