<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>LLM Mechanics — News</title><link>https://llmmechanics.com/en/news/</link><description>New models, research and tools: what’s new in artificial intelligence, how it works and how it is used.</description><language>en</language><item><title>Claude Sonnet 5.5: Anthropic focuses on speed</title><link>https://llmmechanics.com/en/news/sonnet-5-5-evaluer-cout/</link><guid isPermaLink="true">https://llmmechanics.com/en/news/sonnet-5-5-evaluer-cout/</guid><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate><description>Anthropic introduces a new Sonnet version, highlighting faster execution and lower token consumption.</description></item><item><title>Mistral joins Firefox Smart Window</title><link>https://llmmechanics.com/en/news/mistral-mozilla-sources/</link><guid isPermaLink="true">https://llmmechanics.com/en/news/mistral-mozilla-sources/</guid><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate><description>Mozilla and Mistral announce a partnership for Smart Window, Firefox’s browsing assistant, currently in beta.</description></item><item><title>Mistral Small 4: how a mixture of experts works</title><link>https://llmmechanics.com/en/news/mistral-small-4-experts/</link><guid isPermaLink="true">https://llmmechanics.com/en/news/mistral-small-4-experts/</guid><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate><description>A look back at Small 4’s architecture: why can a model contain many experts without activating them all for every token?</description></item></channel></rss>