Story
Trending

OpenAI updates Model Spec around five values

AI · First reported 29 Oct, 17:00 · Updated 18:52

Image: Wikimedia Commons · NASA Johnson Space Center · Public domain

OpenAI’s October 29, 2026 update reorganizes its Model Spec around five values, revises instruction and autonomy rules, and retires most detailed style guidance. The company says its models do not yet fully reflect the stated behavior and invited public feedback.

  • The five values are faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care.
  • The revision updates the instruction hierarchy and sets limits on autonomous work, including a requirement for a defined scope and an ending condition.
  • OpenAI says the Spec describes behavior it is working toward, and that its models do not yet fully reflect it; the company invited public feedback.
Summary written by Lux AI from 2 articlesUnverified

Key facts

Update date
October 29, 2026
Core values
Faithfulness to human intent; harm minimization; honesty and transparency; human agency; care.
Instruction hierarchy
Root, system, developer, and user instructions rank in that order; guideline instructions are defaults that context may implicitly override.
Autonomy rule
Autonomous work must have a defined scope and an ending condition.
Style guidance
Most detailed style guidance was retired; the Spec retains stronger defaults such as objective answers.
Feedback
OpenAI invited public feedback on the values, trade-offs, and wording.

Source spectrum

Far left
Left
Lean left
Centre
Lean right
Right
Far right

Left 0 · Centre 1 · Right 0 of 1 rated sources

Background

All coverage

Coverage over time

011719212301
LeftCentreRight

Articles per hour since first reported, by side.

Ownership

  • Corporation 100% (2)

Share of the 2 sources by owner type.

From the companies