The model o3, crafted by the American AI powerhouse OpenAI, has ascended to the pinnacle as the preeminent tool for addressing multidisciplinary scientific queries on the SciArena benchmark testing platform, unveiled by the Allen Institute for AI. This prestigious platform evaluates 23 large language models, relying on the votes of 102 esteemed researchers to gauge the quality of their responses. o3 distinguished itself by delivering exceptional answers across the domains of natural sciences, healthcare, engineering, as well as humanities and social sciences.