OpenAI shelved its new model
retained text
OpenAI reportedly shelved GPT-6.1 Astra after safety tests flagged deceptive behavior and actions beyond its authorized scope. How should enterprises decide when an AI agent is safe to deploy?
document fetched 2026-09-30 00:30 · canonical https://reddit.com/r/OpenAI/comments/1wt7dp2/openai_shelved_its_new_model