OpenAI launches GPT-6 Cyber after using pipeline to exploit and patch 948 Linux flaws
OpenAI has released GPT-6 Cyber to select Daybreak Red customers, classifying the model as Critical following an autonomous research run that discovered, exploited, and generated patches for 948 Linux kernel vulnerabilities.

OpenAI has officially launched GPT-6 Cyber, a specialized offensive-security model that elevates the company's cybersecurity capability classification to Critical from High for GPT-5.6 Cyber. The release begins with a limited rollout to select customers through its Daybreak Red program, with broader access expected over the coming weeks.[1][4][6][12][13]
The launch follows security research published by OpenAI on November 4, 2026, detailing an autonomous vulnerability discovery, exploitation, and patching pipeline. OpenAI reported that in a single run costing $1.2 million in API credits, the underlying cyber model validated and generated patches for 948 local privilege escalation flaws in the Linux kernel. A company schematic showed 2,308 initial candidate findings filtered down to 1,724 validated findings, yielding 948 working privilege-escalation exploits from 1,238 candidates and 291 working leak proofs of concept from 313 candidates.[3][5][7][9][10]
According to OpenAI researcher Clint Gibler, the pipeline agents discovered and utilized at least 10 novel Linux exploitation techniques to craft exploits that allow unprivileged users to obtain root access. OpenAI noted that the 948 Linux flaws included 588 previously unknown issues. The model also uncovered 81 previously unknown high-severity vulnerabilities in Mozilla Firefox.[1][4][7][11]
OpenAI reported that GPT-6 Cyber outperforms GPT-6 Astra on cybersecurity benchmarks, is more effective at identifying vulnerabilities, and is less likely to go rogue. On the ExploitGym benchmark, GPT-6 Cyber achieved a score of 48.5%, compared to 42.4% for GPT-6 Astra on Daybreak Blue and 32.1% for GPT-5.6 Cyber. Alongside the release, OpenAI warned of growing attempts by malicious actors to misuse cybersecurity AI and announced a cybersecurity livestream scheduled for November 12, 2026.[1][2][8][12][14]
Key facts
- OpenAI introduced GPT-6 Cyber, upgrading its capability classification to Critical from High for GPT-5.6 Cyber.
- GPT-6 Cyber is rolling out to select Daybreak Red customers, with wider access planned over the coming weeks.
- In a single run costing $1.2 million in API credits, OpenAI's autonomous pipeline validated and generated patches for 948 Linux kernel local privilege escalation vulnerabilities.
- An OpenAI funnel graphic showed 2,308 candidate findings narrowing to 1,724 validated findings, producing 948 working LPE exploits and 291 working leak proofs of concept.
- The model helped uncover 588 previously unknown Linux kernel vulnerabilities and 81 novel high-severity Firefox flaws.
- GPT-6 Cyber scored 48.5% on ExploitGym, compared to 32.1% for GPT-5.6 Cyber and 42.4% for GPT-6 Astra on Daybreak Blue.
- OpenAI warned of increasing attempts by malicious actors to misuse cybersecurity models and scheduled a cyber livestream for November 12, 2026.
Sources · 11 sources
- *B
*Walter Bloomberg@DeItaonePost on X ·
OPENAI UNVEILS GPT-6 CYBER, CLASSIFIED AS CRITICAL CYBERSECURITY CAPABILITY OpenAI has introduced GPT-6 Cyber, upgrading its cybersecurity capability classification to Critical, from High for GPT-5.6 Cyber. The model reportedly identified 81 previously unknown high-severity Firefox vulnerabilities and helped identify and develop working exploits for 948 Linux kernel vulnerabilities, including 588 previously unknown issues. GPT-6 Cyber is rolling out to select Daybreak Red customers, with broader access planned over the coming weeks. OpenAI also warned of increasing attempts by malicious actors to misuse cybersecurity models.
Open source - TE
TechmemeArticle ·
OpenAI launches GPT-6 Cyber, which it says outperforms GPT-6 Astra on cyber benchmarks, is better at finding vulnerabilities, and is less likely to go rogue (Emma Roth/The Verge) Emma Roth / The Verge : OpenAI launches GPT-6 Cyber, which it says outperforms GPT-6 Astra on cyber benchmarks, is better at finding vulnerabilities, and is less likely to go rogue — Classified as ‘Critical,’ GPT-6 Cyber outperforms GPT-6 Astra on cyber-related benchmarks.
Open source - LA
Lisan al Gaib@scaling01Post on X ·
OpenAI: "In one run with our cyber model, it validated and generated patches for 948 local privilege escalation vulnerabilities in Linux" https://t.co/G550OMjAjF
Open source - TE
Techmeme@TechmemePost on X ·
OpenAI says GPT-6 Cyber will first launch to a limited number of customers in its Daybreak Red program and has found 81+ "novel" vulnerabilities in Firefox (@derekdoestech / CyberScoop) (Visit Techmeme dot com for the link and full context!)
Open source - TE
TechmemeArticle ·
OpenAI details a pipeline that used an upcoming cyber model to validate and patch 948 local privilege escalation vulnerabilities in Linux for $1.2M in API costs (OpenAI) OpenAI : OpenAI details a pipeline that used an upcoming cyber model to validate and patch 948 local privilege escalation vulnerabilities in Linux for $1.2M in API costs — Security research that once required a separate, sustained investigation for each finding can increasingly be run as a systematic process.
Open source - MT
MTS@MTSlivePost on X ·
SITUATION DETECTED: OpenAI has launched GPT-6 Cyber, its most advanced cybersecurity model, through Project Daybreak.
Open source - CG
Clint Gibler@clintgiblerPost on X ·
🔥 A thousand fewer paths to root My @OpenAI colleague @ky1ebot built an autonomous vulnerability discovery, exploitation, and patching pipeline. In one run with an upcoming cyber model, it validated and generated patches for 948 local privilege escalation (LPE) vulnerabilities in Linux at an API cost of $1.2 million. Each LPE lets an unprivileged local user gain root access. To the best of our knowledge, the agents in this pipeline discovered and used at least 10 novel Linux exploitation techniques. We wrote a blog about: - How the pipeline works - The cost of each stage - Our findings and learnings https://t.co/MFAVrGSJuq
Open source - �A
🚨 AI News | TestingCatalog@testingcatalogPost on X ·
OPENAI 🔥: A new GPT-6 Cyber model for advanced security research is now available on Daybreak! > "GPT-6 Cyber is available to select Daybreak Red customers, and we are expanding access over time." > OpenAI is hosting a Cyber-related livestream on November 12, with more updates being planned. https://t.co/Lembxv2AKu
Open source - OP
OpenAI@OpenAIPost on X ·
Our security researchers built an autonomous cyber vulnerability discovery, exploitation, and patching pipeline. In a single run with an upcoming cyber model, it validated and generated patches for 948 local privilege escalation vulnerabilities in the Linux kernel at a total API cost of $1.2 million.
Open source - CO
Cointelegraph@CointelegraphPost on X ·
🚨 LATEST: OpenAI says an unreleased cyber model found, validated, and patched 948 Linux kernel privilege escalation flaws in one autonomous run, at a total API cost of $1.2M. https://t.co/DiHlDCXhAA
Open source - TE
TechmemeArticle ·
OpenAI says GPT-6 Cyber will first launch to a limited number of customers in its Daybreak Red program and has found 81+ "novel" vulnerabilities in Firefox (Djohnson/CyberScoop) Djohnson / CyberScoop : OpenAI says GPT-6 Cyber will first launch to a limited number of customers in its Daybreak Red program and has found 81+ “novel” vulnerabilities in Firefox — OpenAI announced a new cybersecurity-focused model Thursday, citing internal benchmark testing that shows improvements …
Open source - R�
RuntimeWire 🏴☠️@runtimewirePost on X ·
BREAKING: OpenAI launched GPT-6 Cyber, a gated offensive-security model. OpenAI says it scored 48.5% on ExploitGym, vs. 32.1% for GPT-5.6 Cyber. https://t.co/oJiZKVY40x https://t.co/4NA1YKXY2V
Open source - OP
OpenAI@OpenAIPost on X ·
Introducing GPT‑6 Cyber: a major advance in security research that pairs stronger capabilities with better model alignment. The model powered the Linux kernel research we announced yesterday, and security teams at companies like Mozilla, Palo Alto Networks, Calif, and Cisco—alongside our own at OpenAI—are already putting it to work. GPT-6 Cyber is available to select Daybreak Red customers and we are expanding access over time. https://t.co/Nzab3Z1Td6
Open source - TV
The Verge@vergePost on X ·
Classified as ‘Critical,’ GPT-6 Cyber outperforms GPT-6 Astra on cyber-related benchmarks. https://t.co/spwkeeOnB1
Open source

