Sam Altman Memo: OpenAI's Next Model Has Finished Pretraining, Safety Team Restructured

Sam Altman confirmed in an internal memo that OpenAI's next flagship model has completed pretraining and entered the post-training phase. The memo also disclosed structural changes to the safety team. The completion of pretraining is a significant milestone, as it marks the point where the model's core capabilities are established before fine-tuning, alignment, and safety evaluation. The safety team restructuring, announced alongside the model update, comes as OpenAI faces ongoing scrutiny over its approach to AI safety research and governance.