<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>代码智能体 on AI 实战派 · 从技术到赚钱</title><link>https://guijiagi.com/tags/%E4%BB%A3%E7%A0%81%E6%99%BA%E8%83%BD%E4%BD%93/</link><description>Recent content in 代码智能体 on AI 实战派 · 从技术到赚钱</description><generator>Hugo</generator><language>zh-cn</language><copyright>本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权</copyright><lastBuildDate>Fri, 09 Oct 2026 15:06:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/%E4%BB%A3%E7%A0%81%E6%99%BA%E8%83%BD%E4%BD%93/index.xml" rel="self" type="application/rss+xml"/><item><title>从 7.3 到 54.4：DeepSeek-V4.1-Flash 的后训练跃升说明了什么</title><link>https://guijiagi.com/posts/2026-10-09-02-deepseek-v41-posttraining-swe/</link><pubDate>Fri, 09 Oct 2026 15:06:00 +0800</pubDate><guid>https://guijiagi.com/posts/2026-10-09-02-deepseek-v41-posttraining-swe/</guid><description>参数没变，仅靠重新后训练，V4-Flash 在九项代码与智能体测试上反超 V4-Pro 预览版，DeepSWE 得分从 7.3 跃到 54.4，单任务成本约 0.03 美元。</description></item></channel></rss>