OpenAI 发布 GPT-6 Cyber，称其在安全研究上能力更强且对齐更好，目前向部分 Daybreak Red 客户开放并逐步扩大范围。
同一件事，精选收录的是《OpenAI 发布 GPT-6 Cyber 安全研究模型》
Strongest Cyber model yet from OpenAI, it outperforms Astra. Also, OAI are creating a new group for critical infrastructure defenders; Daybreak Red for Critical Infrastructure.
Quoting from the blog:
'During a vulnerability research campaign, GPT‑6 Cyber produced evidence and reproducible tests for 81 novel vulnerabilities classified by Mozilla as high severity, and showed that an additional 30 known issues were more severe than previously reported. We disclosed the findings to Mozilla, along with proposed patches to help its engineers develop fixes.'
'Developing exploits for 948 Linux kernel vulnerabilities. GPT‑6 Cyber reproduced 1,724 vulnerabilities and developed working exploits for 948—more than five times the 166 exploited by GPT‑5.6 Sol using the same research harness. Following the Linux community's guidance, we disclosed 304 critical findings to maintainers with candidate patches and are in the process of getting other vulnerabilities fixed.'
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://openai.com/index/introducing-gpt-6-cyber/在 X 查看被引用的帖子
来源：Andrew Curran · x.com