OpenAI published a revised Model Spec on October 29, 2026, reorganizing the document that governs how its models behave around five explicit values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care1,2. The update also adds guidance on context, oversight, and agency while retiring most of the Spec's detailed style guidance.
OpenAI described the five values as commitments already reflected across the prior document, now elevated into an organizing framework. Faithfulness to human intent sits at the foundation: OpenAI writes the Model Spec and sets boundaries, developers direct applications within those boundaries, and users direct the model's work subject to both layers of constraints.
The remaining four values carry specific definitions. Minimizing harm means respecting firm safety boundaries and accounting for the foreseeable consequences of a model's actions. Honesty and transparency require the model to be truthful about what it knows, does, and is uncertain about, so that people can understand and evaluate its reasoning. Facilitating human agency means strengthening people's ability to understand, choose, and act for themselves. Care means attending to the circumstances of those affected, following relevant social and professional norms, and producing work people can rely on.
The revision strips out most of the Model Spec's earlier detailed style guidance.
ANALYSIS Consolidating behavioral expectations into five high-level values while removing granular style rules shifts the document from a prescriptive style sheet toward a principles-based framework, a structure more likely to remain stable as OpenAI's model lineup expands.
The update arrives the same day as other major AI-industry developments, including Microsoft's first dollar-denominated Azure revenue disclosure[1] and Google DeepMind's release of cybersecurity-focused Gemini 4 Carbon benchmarks[2].