AI21 Receives a Top Score on Stanford University’s Transparency Index

AI21 Labs received the highest transparency score among commercial general-purpose LLMs on Stanford University’s prestigious Foundation Model Transparency Index (FMTI). The company's score jumped from 25 to 75 in six months, placing it above competitors like OpenAI, Anthropic, and Meta. This high score reflects AI21's increased disclosure around data sourcing, compute, model development, and usage policies, which are critical for enterprise clients in regulated industries. The company connects this commitment to transparency to its products, such as the Jamba-Instruct model with its large 256K context window and Contextual Answers, which is designed to avoid hallucinations by grounding outputs in provided data.