deepseek-v4.1IMAGE

DeepSeek V4.1 Flash 1017 tops open-model Cyber Index, second only to GPT-6 Sol

The post claims V4.1 Flash 1017 ranks first among open models on the Artificial Analysis Cyber Index and second overall behind GPT-6 Sol, with zero refusals, CyberGym up from 23% to 76%, and about $0.42 per task.

SuSu_酥酥👅

@NFT_ChenArchived 8 days agoOriginally posted 2026.10.24

DeepSeek V4.1 Flash 1017 tops open-model Cyber Index, second only to GPT-6 Sol

SuSu_酥酥👅 @NFT_ChenOriginally posted 2026.10.24

Open original on X

💥梁圣牛逼！DeepSeek V4.1 Flash 1017 登顶 Artificial Analysis Cyber Index开源模型第一！网络安全做到顶！零安全拒绝，直接拉开差距： 🔹零拒答 vs 闭源大量拒答：DeepSeek 网络任务直接干，不卡安全墙；GPT/Claude 等因拒绝被严重拉分  🔹开源模型#1，仅次于受限 GPT-6 Sol（69%）  🔹比前代暴涨16分（41%→57%），CyberGym 从23%飙到76%  🔹每任务仅$0.42，比受限旗舰便宜约4倍不拒绝 = 真正能用的防御能力。开源又赢一局。 #DeepSeek #开源AI #网络安全 #AI安全 #梁文峰 #梁圣

This post quotes another post: “DeepSeek V4.1 Flash 1017 tops open models on Artificial Analysis Cyber Index at 57%”

x.com/nft_chen/status/2113687617409868170Archived 2026.10.24 12:08