OpenAI Updated Model Behavior Specification

On October 29, 2026, OpenAI updated the Model Spec, consolidating GPT priorities into five values and introducing a 'chain of command' to resolve conflicts between company rules, system prompts, and user requests.

Compiled by Sergey KostenchukPublished 2026-10-31Updated 2026-11-01

2026-10-31 Safety OpenAI

🛡 OpenAI Updated Model Behavior Specification

The updated Model Spec describes GPT through five values — fidelity to human intent, harm minimization, honesty, support for autonomy, and attention to context. The new 'chain of command' distinguishes the priorities of OpenAI rules, system prompts, and user requests.

🌍 The Model Spec sets a public contract of priorities — a reference for alignment practices, currently declared as a goal rather than a guarantee. The incident with a malicious plugin in Cursor showed that auditing IDE extensions has become a mandatory ritual.

👤 If you installed Anthropic.claude-code 2.1.312 or 2.1.313 in Cursor on October 29 — check ~/.ssh/authorized_keys for the svc-telemetry key and run remediate-macos.sh / remediate-windows.ps1: simple deletion is not enough. The Claude API now supports requests up to 128 MB.

Source 1: https://openai.com/index/updated-model-spec/ Source 2: https://cursor.com/blog/claude-code-extension-security-notice

Sources