Event ·Multi-source

OpenAI 用前沿网络安全模型一周发现约 1000 个 Linux 内核漏洞

OpenAI 的 Eric Wallace 介绍，团队正将大量前沿算力投入开源代码库漏洞挖掘与修复，全部免费。以 Linux 内核为例，传统安全研究人员一年通常发现约 20 个本地提权（LPE）漏洞，而 OpenAI 即将推出的前沿网络安全模型在一周内用约 100 万美元的 token 找到了约 1000 个 LPE 漏洞。过程中该模型还发现了至少 10 类新的 Linux 漏洞利用技术。

Current read

OpenAI 拿自家前沿模型扫 Linux 内核，一周找到约 1000 个提权漏洞，人类一年才 20 个，这个数字对比很直观。

11 sources51° AI heatLast updated 2026/11/4 23:54:59

Evidence

11 sources

Primary sourceGreg Brockman

2026/11/4 23:54:59View source

Related source 1OpenAI Blog

2026/11/4 08:00:00View source

Related source 2@atomic_chat_hq

2026/11/4 17:53:37View source

Related source 3IT之家

2026/11/4 22:51:59View source

Related source 4@OpenAIDevs

2026/11/3 00:09:30View source

Related source 5TestingCatalog

2026/11/3 07:41:31View source

Related source 6rohanpaul_ai

2026/11/3 21:35:42View source

Related source 7AI Will

2026/11/4 07:13:31View source

Related source 8The Information

2026/11/4 14:30:05View source

Related source 9宝玉

2026/11/4 23:46:41View source

Related source 10shao__meng

2026/11/3 01:42:00View 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