<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><title>France's Mistral announces new AI model — Live Feed</title><link>https://www.live-feeds.com/feed/france-s-mistral-announces-new-ai-model</link><atom:link xmlns:atom="http://www.w3.org/2005/Atom" href="https://www.live-feeds.com/feed/france-s-mistral-announces-new-ai-model/rss.xml" rel="self" type="application/rss+xml"/><description>Continuously updated, source-cited coverage.</description>
<item><title>French AI Startup Mistral Unveils New Large 4 Model</title><link>https://www.live-feeds.com/feed/france-s-mistral-announces-new-ai-model</link><guid isPermaLink="false">https://www.live-feeds.com/feed/france-s-mistral-announces-new-ai-model#u132292</guid><pubDate>Sun, 18 Oct 2026 15:25:16 +0000</pubDate><description>French artificial intelligence startup Mistral unveiled its largest open-weight model on Tuesday, ending a five-month gap since its previous major release. Known as Mistral Large 4, Le Chonk, or ML4, the one-trillion-parameter text output model is designed for clients seeking to keep data private on their own servers rather than relying on closed systems from American competitors like Anthropic and OpenAI. Chief Executive Officer Arthur Mensch stated at a conference in Abu Dhabi that the new product outperforms certain Chinese open-weight models, including in cybersecurity aspects. The release</description></item>
</channel></rss>