<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>ECI on Text Matrix</title><link>https://txtmix.com/tags/eci/</link><description>Recent content in ECI on Text Matrix</description><generator>Hugo</generator><language>zh-cn</language><lastBuildDate>Tue, 21 Jul 2026 20:06:14 +0800</lastBuildDate><atom:link href="https://txtmix.com/tags/eci/index.xml" rel="self" type="application/rss+xml"/><item><title>中国 AI 模型是否追上了美国前沿？Kimi-K3 深度解读</title><link>https://txtmix.com/posts/tech/have-chinese-ai-models-caught-up-translation/</link><pubDate>Mon, 20 Jul 2026 21:55:41 +0800</pubDate><guid>https://txtmix.com/posts/tech/have-chinese-ai-models-caught-up-translation/</guid><description>&lt;blockquote>
&lt;p>原文：&lt;a href="https://scaling01.substack.com/p/have-chinese-ai-models-caught-up" target="_blank" rel="noopener noreffer ">Have Chinese AI Models Caught Up to the US Frontier?&lt;/a>（作者 Zito / scaling01，2026-07-19）&lt;/p>
&lt;p>译者按：MoonshotAI（月之暗面）发布 Kimi-K3 时，国内同行振奋、海外也热闹了一周。但 scaling01 这篇长文给出了一份冷静的反向论证——把 Artificial Analysis Intelligence Index（AAI）与 EpochAI Capability Index（ECI）两套评测对照读后，结论是：表层 benchmark 上 Kimi-K3 已经逼近甚至超过美国前沿；可一旦换用更严谨的 ECI，中美之间的真实差距仍在 4 到 5 个月以上，前看差距更是接近一年。&amp;ldquo;追上&amp;quot;二字，言之过早。&lt;/p></description></item></channel></rss>