<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
  <title>MATRIX · Nanbeige4.3-33B</title>
  <link>https://derekwang2002.github.io/ai-intelligence/en/projects/nanbeige-4-3-33b/</link>
  <description>Material changes in AI models, agents, engineering and infrastructure, with primary sources and recommendations.</description>
  <language>en</language>
  <lastBuildDate>Mon, 02 Nov 2026 16:56:00 GMT</lastBuildDate>
  <atom:link href="https://derekwang2002.github.io/ai-intelligence/en/feeds/projects/nanbeige-4-3-33b.xml" rel="self" type="application/rss+xml"/>
  <item>
    <title>Nanbeige releases an open 33B model with a 512K context window</title>
    <link>https://derekwang2002.github.io/ai-intelligence/en/events/ev-20261024-02/</link>
    <guid isPermaLink="false">ev-20261024-02-published</guid>
    <pubDate>Sun, 25 Oct 2026 15:57:07 GMT</pubDate>
    <category>New</category>
    <category>TRIAL</category>
    <description>&lt;p&gt;Nanbeige released Apache-2.0 weights for Nanbeige4.3-33B with a 512K context window and deployment paths for Transformers, SGLang and vLLM. The model card reports coding and office-task results, but evaluation is publisher-only and adoption is still very early.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Why read: &lt;/strong&gt;Open 33B/512K weights add a local long-context option whose cost and effective length still need testing.&lt;/p&gt;&lt;p&gt;Recommendation: TRIAL&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://huggingface.co/Nanbeige/Nanbeige4.3-33B&quot;&gt;Nanbeige4.3-33B model card&lt;/a&gt;&lt;/p&gt;</description>
  </item>
</channel>
</rss>
