In a twist of irony, OpenAI announced its upcoming GPT-5.5 Cyber tool will feature restricted access. The rollout will be limited exclusively to "critical cyber defenders," mirroring the very approach OpenAI recently criticized Anthropic for using with Mythos.