Event ·Multi-source

DeepSeek V4.1 Flash 1017 登顶开放模型网络防御榜单

DeepSeek 发布的 DeepSeek V4.1 Flash 1017 在 Artificial Analysis Cyber Index 中得分 57%,成为开放可用模型中排名第一,总榜第二。该指数由 CWE-Bench-AA、DeepsecBench-AA 和 CyberGym-E2E-AA 三个基准组成,评测智能体网络防御能力。相比前代 DeepSeek V4.1 Flash 的 41%,此次提升 16 点,其中 CyberGym-E2E-AA 从 23% 升至 76%。总榜第一是 OpenAI 受信任访问计划下的 GPT-6 Sol(Daybreak Blue, max),得分 69%,但公共版本因安全拒绝拉低了成绩。DeepSeek V4.1 Flash 1017 每任务成本 $0.42,比 GPT-6 Sol 便宜 4 倍。

Current read

DeepSeek 新版开源模型在网络防御榜单上拿了开放模型第一,总分只差 GPT-6 受限版 12 分,价格还便宜 4 倍。

11 sources74° AI heatLast updated 2026/10/23 17:23:09

Evidence

11 sources
Primary sourceArtificial Analysis
View source
Related source 1TestingCatalog
View source
Related source 2karminski-牙医 (AI工具)
View source
Related source 3Teortaxes
View source
Related source 5Latent Space (swyx)
View source
Related source 7techcrunch
View source
Related source 8Bindu Reddy
View source
Related source 9kimmonismus
View source
Related source 10theverge
View 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.

DeepSeek V4.1 Flash 1017 登顶开放模型网络防御榜单 · 事件专题 | AITOP