<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Deepseek on Ask Sendai</title>
    <link>https://asksendai.com/tags/deepseek/</link>
    <description>Recent content in Deepseek on Ask Sendai</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Wed, 16 Sep 2026 23:05:21 +0000</lastBuildDate>
    <atom:link href="https://asksendai.com/tags/deepseek/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Enclave: DeepSeek V4.1 Flash is Now Our Best Hacking Model</title>
      <link>https://asksendai.com/post/2026/05/16/enclave--d/</link>
      <pubDate>Wed, 16 Sep 2026 23:05:21 +0000</pubDate>
      <guid>https://asksendai.com/post/2026/05/16/enclave--d/</guid>
      <description>&lt;figure&gt;&lt;img src=&#34;https://www.datocms-assets.com/197073/1789462052-deepseek-v4-1-flash-is-now-our-best-hacking-model.jpg&#34;&#xA;    alt=&#34;Enclave: DeepSeek V4.1 Flash is Now Our Best Hacking Model&#34;&gt;&#xA;&lt;/figure&gt;&#xA;&#xA;&lt;p&gt;&#xA;DeepSeek’s 11/11 result showed why advanced agent benchmarks need to check both the outcome and the attack path: our audit confirmed six planned exploits and found five unexpected routes. Across the full benchmark, the model used 2,349 Bash commands and almost two hours and 38 minutes of active model time. The median successful run took four minutes and 38 seconds. The provider reported 268.3 million input tokens and about two million output tokens.&#xA;&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
