<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
  <title>MATRIX · WeVisOne-VL</title>
  <link>https://derekwang2002.github.io/ai-intelligence/en/projects/wevisone-vl/</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/wevisone-vl.xml" rel="self" type="application/rss+xml"/>
  <item>
    <title>Tencent releases WeVisOne-VL weights for unified visual perception</title>
    <link>https://derekwang2002.github.io/ai-intelligence/en/events/ev-20261023-01/</link>
    <guid isPermaLink="false">ev-20261023-01-published</guid>
    <pubDate>Fri, 23 Oct 2026 16:04:55 GMT</pubDate>
    <category>New</category>
    <category>TRIAL</category>
    <description>&lt;p&gt;Tencent released WeVisOne-VL 2B and 4B model weights plus a required 9.6M-parameter pixel decoder for detection, referring localization, counting, point grounding, OCR, segmentation and alpha matting. The Hugging Face repositories are ungated and list Apache-2.0, while the GitHub detector does not expose a detected license. Published benchmarks are vendor-reported.&lt;/p&gt;&lt;p&gt;Recommendation: TRIAL&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://github.com/Tencent/WeVisOne-VL&quot;&gt;Tencent WeVisOne-VL&lt;/a&gt; · &lt;a href=&quot;https://huggingface.co/tencent/WeVisOne-VL-4B&quot;&gt;WeVisOne-VL-4B&lt;/a&gt; · &lt;a href=&quot;https://huggingface.co/tencent/WeVisOne-VL-2B&quot;&gt;WeVisOne-VL-2B&lt;/a&gt;&lt;/p&gt;</description>
  </item>
</channel>
</rss>
