Event ·Multi-source

DeepSeek 与 Claude Opus 找漏洞能力相当，补漏洞可信度却不同

博主 teortaxesTex 观察到一个攻防不对称现象：DeepSeek 与 Anthropic 的 Astra/Opus 在发现漏洞上表现接近，但面对 20 年历史的企业级软件栈时，他认为 DeepSeek 修补漏洞的可信度不足。他据此提出问题：这种不对称会如何影响安全场景下的模型需求与算力分配。该观点以推文形式发布，未附基准数据，属于个人经验判断。

Current read

一个搞安全的推友观察到：找漏洞 DeepSeek 和 Opus 打平，但修老系统还得靠 Opus，聊聊算力怎么分配。

11 sources48° AI heatLast updated 2026/10/23 11:15:12

Evidence

11 sources

Primary sourceTeortaxes

2026/10/23 11:15:12View source

Related source 1The Information

2026/10/21 13:00:16View source

Related source 2rohanpaul_ai

2026/10/21 19:48:33View source

Related source 3TestingCatalog

2026/10/22 08:23:08View source

Related source 4DAIR.AI

2026/10/22 14:24:02View source

Related source 5Artificial Analysis

2026/10/22 02:04:56View source

Related source 6roon

2026/10/22 03:58:10View source

Related source 7Andrew Curran

2026/10/22 13:53:03View source

Related source 8@koltregaskes

2026/10/21 11:34:00View source

Related source 9Decoder

2026/10/21 12:34:57View source

Related source 10IT之家

2026/10/21 12:42:51View 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