The system serves authorized teams by automating vulnerability discovery and evaluating exploit feasibility across major platforms. During internal trials, researchers deployed the model to uncover dozens of high-severity flaws in Mozilla Firefox and generate working exploits across hundreds of Linux vulnerabilities. Access remains strictly gated behind a dedicated security tier to limit misuse while system maintainers receive suggested remediation patches.
Key points
Access to the new model is currently restricted to vetted users operating within the spec…: Access to the new model is currently restricted to vetted users operating within the specialized Daybreak Red service tier.
OpenAI plans to broaden the critical model's deployment to all approved Daybreak Red cust…: OpenAI plans to broaden the critical model's deployment to all approved Daybreak Red customers over the coming weeks.
The earlier GPT-5.6 Cyber model, carrying a High safety rating, has been opened to all us…: The earlier GPT-5.6 Cyber model, carrying a High safety rating, has been opened to all users within the Daybreak Red tier.
Sources checked by our analysis pipeline: 1
Nov 5 : OpenAI released GPT-6 Cyber, a model built to find and exploit security flaws, to a small group of vetted customers on Thursday, rating it "Critical" in cybersecurity under its safety framework.
The model is meant to help security researchers find vulnerabilities, test how they could be exploited and fix flaws before attackers do, OpenAI said, amid increasing attempts to misuse AI by malicious cyber actors.
Here are more details:
• OpenAI said its researchers used the model to find 81 previously unknown Firefox vulnerabilities that Mozilla classified as high severity. Mozilla has begun releasing fixes
• It also helped develop working exploits for 948 Linux kernel vulnerabilities, including 588 that were previously unknown, OpenAI said
• The company has also disclosed 304 critical Linux findings to maintainers, along with proposed patches
• OpenAI has raised its cybersecurity classification for GPT-6 Cyber to "Critical" under its safety framework, from "High" for GPT-5.6 Cyber
• The company said it is seeing more attempts to misuse AI, including by malicious cyber actors and groups seeking to distill its models
Google is facing claims that a student used an artificial intelligence chat service on a school laptop in the United States to plan a suicide attempt. Details regarding the specific AI tool involved and the school district have not been disclosed.
The legal action accuses the tech giant of aggressively pushing artificial intelligence into educational settings without disclosing the severe risks to young users. Filed by parents in California and relocated to federal court, the suit seeks monetary damages following an incident involving school-issued technology. The case intensifies scrutiny over Big Tech's deployment of generative algorithms in classrooms without adequate safeguards for vulnerable children.
Rivet, a startup in the Y Combinator Fall 2025 batch, has launched an open-source orchestrator designed for scaling and managing agentic workloads. Built on the actor model, the system assigns each agent a lightweight "Rivet Actor" backed by S3 rather than provisioning a heavyweight sandbox. The self-hostable technology runs on plain Node.js, Bun, or Rust, allowing developers to deploy the control plane locally, on Docker, on Kubernetes, or on virtual machines within their own cloud infrastructure.