According to OpenAI, on October 29, 2026, the company updated its public Model Spec, a document outlining how its GPT models should behave as they undertake more consequential tasks. The revised specification is built around five core values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care.
The update adds guidance on context, oversight, and resolving conflicting instructions, while removing most detailed style guidance. OpenAI noted that its models do not yet fully reflect the updated spec.