I think Anthropic went a lot further with the marketing around Mythos/Glasswing.

OpenAI talked a lot about potential future risks. Anthropic went around saying "this model is too dangerous to give to people at all".