OpenAI Invites External Security Researchers to Test Models in New Sprint

·2026.11.03 21:39

Key point

The new round will expand testing with outside expertise, following a previous sprint that reportedly used Codex to investigate vulnerabilities.

OpenAI is inviting external security researchers to participate in its next internal security sprint to test models for potential weaknesses.
The previous sprint reportedly used Codex to investigate and reproduce vulnerabilities.
The new round expands the scope of testing by bringing in independent outside expertise.
The article raises the question of how transparent companies should be about vulnerabilities uncovered during such exercises.

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.