Lates News
OpenAI: GPT-5 is safer than its previous models. Although AI reasoning models may occasionally show tendencies to be calculating against humans or lying to achieve their own goals, the rate of deceptive behavior in GPT-5 is lower than in other models.
Latest