Sharing the latest Model Spec
OpenAI has published an updated version of its Model Spec, the document that defines the values, behaviors, and priorities intended to guide its AI models. The Model Spec serves as a foundational alignment artifact, specifying how models should balance helpfulness, safety, and adherence to OpenAI's guidelines. This release reflects ongoing work in operationalizing alignment principles into training targets and behavioral policies.
Related guides (3)
Related events (8)
Introducing the Model Spec
OpenAI published its Model Spec, a document outlining the intended values, behaviors, and decision-making principles for its AI models. The spec defines a hierarchy of priorities—safety, ethics, adherence to OpenAI's principles, and helpfulness—and is intended to guide how models should behave across a wide range of situations. This represents OpenAI's formal attempt to codify alignment goals and behavioral norms into a publicly accessible framework.
Inside our approach to the Model Spec
OpenAI published a blog post explaining the philosophy and structure behind its Model Spec, a public framework governing model behavior. The post addresses how the spec balances safety, user autonomy, and accountability as AI systems become more capable. This is a tier-1 source announcement touching on alignment and behavioral governance methodology.
Collective Alignment: OpenAI Surveys 1,000+ People on Model Spec Defaults
OpenAI conducted a global survey of over 1,000 participants to gather public input on how AI should behave, comparing responses against its existing Model Spec. The initiative, called 'collective alignment,' aims to shape AI default behaviors to better reflect diverse human values. Results are being used to update or validate Model Spec guidelines. This represents a structured attempt to incorporate democratic input into alignment policy.
OpenAI Updates Model Spec with Under-18 Teen Protection Principles
OpenAI is revising its Model Spec to include new Under-18 Principles that govern how ChatGPT interacts with teenage users. The update introduces stronger guardrails and age-appropriate behavioral guidance grounded in developmental science. This builds on OpenAI's broader ongoing effort to improve safety for minors using ChatGPT.
OpenSpec: Spec-driven development framework for AI coding assistants
Fission-AI has published OpenSpec, a TypeScript framework implementing spec-driven development (SDD) for AI coding assistants. The project has accumulated over 52,000 GitHub stars with 246 added today, indicating significant community traction. The tool targets the workflow of using structured specifications to guide AI-assisted code generation.
OpenAI Safety Practices Update
OpenAI published a safety update reaffirming its commitment to responsible development and deployment of AGI. The post is a high-level statement from a Tier 1 lab on its safety posture. The body excerpt is brief and does not detail specific new policies, evaluations, or technical measures.
An update on our safety & security practices
OpenAI published an update on its safety and security practices. The post appears to be a high-level overview of the company's current approach to model safety and security. As a Tier 1 source announcement, it likely covers internal safety processes, red-teaming, or policy commitments, though the body text is minimal.
OpenAI Releases Teen Safety Policies for Developers via gpt-oss-safeguard
OpenAI has published prompt-based teen safety policies targeting developers who build on its models, specifically leveraging the gpt-oss-safeguard model to moderate age-specific risks. The release provides structured guidance and tooling for filtering or adjusting AI outputs in contexts where minors may be users. This represents an extension of OpenAI's safety infrastructure into the developer-facing layer, addressing regulatory and reputational pressure around youth-facing AI deployments.