Introducing Trusted Access for Cyber

February 5, 2026

TL;DR

GPT-5.3-Codex is OpenAI's most cyber-capable frontier reasoning model, capable of complex autonomous tasks.
Cybersecurity advancements with these models can strengthen the ecosystem and introduce new risks.
Trusted Access for Cyber is a pilot program using an identity and trust-based framework for controlled access to enhanced cyber capabilities.
The goal is to accelerate cyber defense by aiding vulnerability discovery and remediation.
OpenAI is committing $10 million in API credits to accelerate cyber defense efforts.
Frontier models have mitigations against clearly malicious requests, and automated monitors will detect suspicious activity.
Users can verify identity at chatgpt.com/cyber, and enterprises can request trusted access via their OpenAI representative.
An invite-only program is available for security researchers needing more permissive models for defensive work.
The Cybersecurity Grant Program will offer API credits to teams with a track record in vulnerability identification and remediation.

Continue reading
the original article