Collective alignment: public input on our Model Spec

OpenAI surveyed over 1,000 people worldwide, compared their preferred model behavior with its Model Spec, and adopted some changes from disagreements. The post says participants ranked 4 completions per prompt, OpenAI compared them with a GPT-5 Thinking-based Model Spec Ranker, and released the dataset on HuggingFace. The key issue is default behavior; the captured post does not disclose the full list of adopted changes.

Why it matters: OpenAI turns >1,000 public preference rankings into Model Spec edits and releases the dataset, so HKR-H/K/R all pass. The real signal is default-behavior governance, but the excerpt does not show the full change list, keeping it in the 78–84 band.