Mission
The mission of our organization is to advance the public understanding and governance of artificial intelligence (AI) by rigorously measuring and publishing scientific evaluations of AI capabilities, alignment, and risks. We serve researchers, policymakers, nonprofit institutions, educators, and the general public by providing high-quality, open-access evaluations, tools, and benchmarks. Our long-term goal is to ensure that rapid AI development proceeds in a manner that is transparent, accountable, and aligned with public benefit. As AI capabilities accelerateparticularly in large language models, robotics, and multi-modal systemsthere is an increasing societal need for neutral, scientific measurement of progress. Unlike corporate leaderboards or proprietary evaluations, our organization will develop public-good infrastructure for measurement: open-weight evaluations, reproducible testbeds, and longitudinal capability tracking. We will develop and publish on benchmarks to track AI capa