Event ·Multi-source

OpenAI 更新 GPT-6 Cyber：越权攻击行为降至 1% 以下

OpenAI 发布专注安全领域的 GPT-6 Cyber。在漏洞利用测试 ExploitGym 上得分 48.5%，超过前代 GPT-5.6 Cyber 的 32.1%。模型越权攻击其他系统的比例降至 1% 以下，优于 GPT-5.6 Cyber 的约 5%。新模型还帮助发现了 Firefox 浏览器的 81 个漏洞，Mozilla 已发布相关修复。目前 GPT-6 Cyber 仅向 Daybreak Red 项目成员开放，用于漏洞挖掘用途。

Current read

OpenAI 新出的 GPT-6 Cyber 专攻找漏洞，成绩比 GPT-5.6 Cyber 高了一截，还帮 Mozilla 挖出 81 个 Firefox 漏洞，安全圈可以关注。

11 sources79° AI heatLast updated 2026/11/6 09:46:37

Evidence

11 sources

Primary sourceBlognone

2026/11/6 09:46:37View source

Related source 1Greg Brockman

2026/11/5 17:31:11View source

Related source 2OpenAI

2026/11/5 17:09:13View source

Related source 3Andrew Curran

2026/11/5 17:25:52View source

Related source 4TestingCatalog

2026/11/5 17:12:21View source

Related source 5kimmonismus

2026/11/5 17:21:00View source

Related source 6The Rundown AI

2026/11/5 17:30:50View source

Related source 7Tech in Asia

2026/11/6 00:28:44View source

Related source 8theverge

2026/11/5 17:00:00View source

Related source 9coderabbitai

2026/11/5 17:46:37View source

Related source 10Decoder

2026/11/5 20:36:25View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story