NIST announced today that it reached an agreement with OpenAI and Anthropic to gain access to both organization's AI models for safety testing, research, and evaluation. You can find NIST's AI Safety Institute announcement here, along with the Verge's reporting on the announcement.