Models & Labs
OpenAI Launches GPT-5.6-Cyber for Security Testing
OpenAIhigh confidence
Why it matters
- →GPT-5.6-Cyber provides a specialized tool for cybersecurity professionals.
- →It enhances capabilities in vulnerability research and exploit validation.
- →This model represents OpenAI's strategic entry into the cybersecurity sector.
OpenAI has launched GPT-5.6-Cyber, a new model specifically designed for cybersecurity purposes. This model is accessible through Daybreak Red and is intended for authorized use in vulnerability research, exploit validation, and security testing. By entering the cybersecurity field, OpenAI aims to provide advanced tools to help professionals better identify and address potential threats. This move highlights the growing role of AI in enhancing cybersecurity measures.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
More in Models & Labs
© The Verge AIModels & Labsother
Google launches Guided Vision in Gemini Live
Google is bringing real-time audio scene description to Android via Gemini Live, directly challenging Apple’s VoiceOver Live Recognition. This feature targets users with low vision by providing immediate audio cues and follow-up Q&A capabilities for physical objects. It integrates deeply into the accessibility ecosystem through TalkBack, moving beyond simple text reading to contextual environmental awareness. The move signals a shift toward multimodal AI as a standard utility for daily navigation rather than just a novelty.
The Verge AI
© TechCrunch AI