Skip to content

Back to the catalog

openai.comarticle

An updated Model Spec for more capable and beneficial AI

OpenAI's 29 October 2026 post announces an updated Model Spec, reorganised around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care. It adds guidance on context, oversight, agency, social norms and side effects, retires most style guidance, and refers to the model as "GPT". OpenAI says its models do not yet fully reflect the Spec.

TL;DR OpenAI reorganises its Model Spec around five core values, adds new defaults on context and agency, and drops most style rules.
Open original ↗

Organisations named here

  • OpenAI
  • Model Spec

Technologies named here

  • GPT
  • LLM
  • Model Spec

Techniques named here

  • Assistance games framing of alignment
  • Chain of command
  • New training methods for value learning
  • Value alignment

Related in ai-safety