OpenAI Says New GPT-6 Cyber Model Hunts Flaws Without Crossing the Line

Rated Cri­ti­cal for cy­ber­se­cu­rity, Ope­nAI’s new model beats GPT-5.6 Cyber at ex­ploi­ta­tion while sta­ying in scope more re­liably. The post Ope­nAI Says New GPT-6 Cyber Model Hunts Flaws Without Cros­sing the Line ap­pea­red first on Se­cu­rity­Week .