As interest and investment grows around AI applications in nuclear power plants, there remains a gap in standardized benchmarks that can quantitatively compare and measure the quality and reliability of new products.

Nuclear-tailored AI developer Atomic Canyon is moving to fill that gap by entering into a new strategic partnership with Idaho National Laboratory to develop and release the “first comprehensive benchmark suite for evaluating retrieval-augmented generation (RAG) and large language models (LLMs) in nuclear applications.”