<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[EXL3]]></title><description><![CDATA[<p dir="auto">Seems to be objectively one of the BEST quantization formats as of now, with the added benefit that the inference engines run much better than llama.cpp (GGUF) for CPU offload.</p>
<p dir="auto">For example, numbers on <a href="https://huggingface.co/turboderp/Qwen3.8-Flash-Next-exl3" rel="nofollow ugc">Qwen3.8-Flash-Next-exl3</a>:</p>
<p dir="auto"><img src="/assets/uploads/files/1790613719744-f70f28de-95b8-4a60-8961-df981b227a6f-image-resized.jpeg" alt="image.jpeg" class="img-markdown  img-fluid" width="2240" height="1600" /></p>
<p dir="auto"><img src="/assets/uploads/files/1790613724890-cebda33e-0bec-4fae-8b37-96e5a5f8fa04-image-resized.jpeg" alt="image.jpeg" class="img-markdown  img-fluid" width="2240" height="1600" /></p>
<p dir="auto"><img src="/assets/uploads/files/1790613730339-e4eb4d23-52f2-43b7-9180-29ceec0c710d-image-resized.jpeg" alt="image.jpeg" class="img-markdown  img-fluid" width="2240" height="1696" /></p>
<div class="airigs-disclosure airigs-disclosure--no" data-ai-disclosure="no"><i class="fa fa-user" aria-hidden="true"></i><span class="airigs-disclosure__label">No AI used</span></div>]]></description><link>https://airigs.net/topic/8/exl3</link><generator>RSS for Node</generator><lastBuildDate>Mon, 02 Nov 2026 07:04:02 GMT</lastBuildDate><atom:link href="https://airigs.net/topic/8.rss" rel="self" type="application/rss+xml"/><pubDate>Mon, 28 Sep 2026 16:42:34 GMT</pubDate><ttl>60</ttl></channel></rss>