The capabilities of LLMs extend into materials science, providing novel insights and task automation. This perspective paper outlines the potential for these models to help researchers navigate the vast landscape of material properties and associated data.

• Discusses the flexibility of LLMs in handling ambiguous requirements.
• Provides case studies on task automation and knowledge extraction.
• Regards LLMs as tools for accelerating cross-domain exploration.
• Emphasizes LLMs as ‘tireless workers’ rather than ‘oracles of novel insight’.

The synthesis of LLMs and materials science heralds a new era of research, potentially making groundbreaking discoveries more accessible and efficient.