<?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[Throughput & Latency]]></title><description><![CDATA[Tokens/s, time-to-first-token, concurrency curves.]]></description><link>https://airigs.net/category/18</link><generator>RSS for Node</generator><lastBuildDate>Mon, 02 Nov 2026 05:55:22 GMT</lastBuildDate><atom:link href="https://airigs.net/category/18.rss" rel="self" type="application/rss+xml"/><pubDate>Mon, 28 Sep 2026 17:10:59 GMT</pubDate><ttl>60</ttl><item><title><![CDATA[Qwen 3.8 Flash Next on RTX 3090's]]></title><description><![CDATA[I FEEL like this is too good to be true, runs a patched vLLM on 2 RTX 3090's:
GitHub recipe
Numbers are absolutely insane.
Something more in-line with what I'd expect is what Minachist achieved:
HF discussion
Blog post
]]></description><link>https://airigs.net/topic/9/qwen-3.8-flash-next-on-rtx-3090-s</link><guid isPermaLink="true">https://airigs.net/topic/9/qwen-3.8-flash-next-on-rtx-3090-s</guid><dc:creator><![CDATA[fedes1to]]></dc:creator><pubDate>Mon, 28 Sep 2026 17:10:59 GMT</pubDate></item></channel></rss>