OpenAI Updates Core Model Values: Fidelity to Human Intent, Harm Minimization, Honesty & Transparency, Human Agency, and Care
On October 29, 2026, OpenAI released an updated Model Spec, restructuring the specification around five core values. The new spec adds guidance on context, supervision, and agency, adjusts some default behaviors, retires numerous detailed style requirements, and states that it is exploring new training methods.
OpenAI Updates Core Values for Its Models: Fidelity to Human Intent, Harm Minimization, Honesty and Transparency, Human Agency, and Care
On October 29, 2026, OpenAI released an updated Model Spec, restructuring the guidelines that describe how its models should behave around five core values: Fidelity to Human Intent, Harm Minimization, Honesty and Transparency, Human Agency, and Care.
The new specification adds guidance on context, oversight, and agency, adjusts some default behavioral settings, and retires a large number of detailed style requirements. OpenAI also stated that it is exploring new training methods to help models understand why these principles are important.
Source: OpenAI's Updated Model Spec