<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:media="http://search.yahoo.com/mrss/">
  <channel>
    <title>Bramble: Open-weight AI models</title>
    <link>https://news.ketelle.co.uk/entities/open-weight-ai-models</link>
    <description>Stories involving Open-weight AI models, told neutrally, with how each side covers them.</description>
    <language>en-gb</language>
    <atom:link href="https://news.ketelle.co.uk/entities/open-weight-ai-models/feed.xml" rel="self" type="application/rss+xml"/>
    <item>
      <title>Mistral releases Large 4 openly, pitching it for cybersecurity use</title>
      <link>https://news.ketelle.co.uk/stories/mistral-releases-large-4-model-publicly-aimed-at-cybersecurity</link>
      <guid isPermaLink="true">https://news.ketelle.co.uk/stories/mistral-releases-large-4-model-publicly-aimed-at-cybersecurity</guid>
      <pubDate>Thu, 29 Oct 2026 14:42:02 +0000</pubDate>
      <category>Technology</category>
      <description>French AI company Mistral made its Large 4 model (nicknamed "Le Chonk") freely downloadable on 29 October, saying it matches or beats leading Chinese open-weight models in coding and cybersecurity. Mistral says open weights protect users from being cut off, while a security expert notes the model could also be abused. (4 outlets: 1 left · 3 centre · 0 right)</description>
      <media:content url="https://dam.mediacorp.sg/image/upload/s--jRS6MPqI--/c_fill,g_auto,h_676,w_1200/fl_relative,g_south_east,l_mediacorp:cna:watermark:2024-04:reuters_1,w_0.1/f_auto,q_auto/v1/one-cms/core/2026-10-29T140336Z_2_LYNXNPEM9S1U2_RTROPTP_3_MISTRAL-CERN.JPG?itok=pY3PAWXI" medium="image"/>
    </item>
  </channel>
</rss>
