---
format: "aidr-story-markdown/v1"
id: "c876496118766282bf936d94e85af52c106c72c558b1b21810f2463c7a54d531"
canonical_url: "https://aidr.today/c8764961?lang=en"
title: "vLLM v0.34.0"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-16T23:22:00.000Z"
category: "Open Source"
topics: ["llm","open-source"]
source_urls: []
summary: "vLLM v0.34.0 includes 1,061 commits and adds 10 new models. The update enables an extensible KV cache by default, removes Model Runner V1, and provides speedups on Blackwell and MI355X hardware. The default block size is now 64."
---

# vLLM v0\.34\.0

> [Open the canonical story](<https://aidr.today/c8764961?lang=en>)

**Published:** 2026-11-16T23:22:00.000Z
**Category:** Open Source
**Topics:** llm, open\-source

## Summary

vLLM v0\.34\.0 includes 1,061 commits and adds 10 new models\. The update enables an extensible KV cache by default, removes Model Runner V1, and provides speedups on Blackwell and MI355X hardware\. The default block size is now 64\.

## Sources

_No safe source URL was recorded._

