vLLM v0.25.0 发布：MRv2 默认、新模型与性能优化

AI 摘要

vLLM 发布 v0.25.0 版本，包含 558 个提交，来自 232 位贡献者。该版本将 Model Runner V2 设为所有稠密模型的默认执行路径，并移除了 PagedAttention 遗留实现。新模型支持包括 LLaVA-OneVision-2、GLM-5、DeepSeek-V3.2 等，并引入了新的流式解析引擎和通用投机解码。性能优化涵盖 GLM-5.2、DeepSeek 和 Blackwell 硬件，显著提升了吞吐量。

正文节选

# vLLM v0.25.0 Release Notes ## Highlights This release features 558 commits from 232 contributors (64 new)! * **Model Runner V2 is now the default for all dense models** (#44443). Building on quantized-model support from the previous release, MRv2 is now the standard execution path, with new support for EVS (#46535), realtime embeddings (#46762), prefix caching for Mamba hybrid models (#42406), multimodal-prefix bidirectional attention (#46942), and dynamic speculative decoding compati

发布时间：2026-07-12 04:06

抓取时间：2026-08-02 00:23

来源机构：vLLM