On March 31, the AI giant published applications for feedback sessions on an upcoming open language model, the second such model since OpenAI’s LLMs went private after GPT-2.