<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Vllm on ArkNill</title><link>https://arknill.github.io/tags/vllm/</link><description>Recent content in Vllm on ArkNill</description><generator>Hugo</generator><language>en</language><lastBuildDate>Sat, 25 Apr 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://arknill.github.io/tags/vllm/index.xml" rel="self" type="application/rss+xml"/><item><title>Quantization, Determinism, and Thinking Tokens: Running Open-Source LLMs in Production</title><link>https://arknill.github.io/blog/quantization-determinism-thinking-production/</link><pubDate>Sat, 25 Apr 2026 00:00:00 +0000</pubDate><guid>https://arknill.github.io/blog/quantization-determinism-thinking-production/</guid><description>FP8 is the production floor. Q4 MoE loses 16% on CJK. vLLM is non-deterministic under MTP. Thinking tokens eat 90% of your budget on the wrong tasks. Hard lessons from operating Qwen 3.5/3.6 35B across 3 nodes.</description></item></channel></rss>