Living definition
What is OpenAI Safety Evaluations Hub?
AI Safety, Eval & AlignmentSecurity & Identity
- As of
- 2026-10-01
- Revision
- 2026-10-06.1
- Method
- v1.4.0
Definition
The term, in context
AI-assisted draft · approved dataset
The Safety Evaluations Hub is a public page on which OpenAI publishes results from its internal safety evaluations of released models, covering areas such as refusal of disallowed content, robustness to jailbreak attempts, factual accuracy, and adherence to the instruction hierarchy. Scores are presented per model and refreshed as new models ship, letting outside readers compare successive releases on the same measures.
Live readiness status
Status as of the current dateline
AI-assisted assembly · derived results
As of 2026-10-01, verified readiness is 25 (claimed 55, reported 35, gap 30) — Strong evidence strength; current signals suggest Use the hub as one input to model due diligence.
The readiness fact belongs to the canonical Readiness Verdict for OpenAI Safety Evaluations Hub.
Dataset approval
Human editorial release
Codex — delegated editorial approver (operator authorization 2026-10-06)
- Dataset
- 2026-10-06.1
- Hash
- sha256:0df9a4b64770fab5c2feaca298223eb10279a31b43487556407f496c9b022685
- Approved
- 2026-10-06