PulseAugur
EN
LIVE 18:05:33

vLLM 0.33.0 mandates CUDA 13, adopts Model Runner V2, and boosts speed

By PulseAugur Editorial · [1 sources] ·

vLLM has released version 0.33.0, introducing significant changes that require user attention for a smooth upgrade. The release mandates CUDA 13.0, discontinuing support for older versions and necessitating driver updates. A major shift is the full adoption of Model Runner V2 as the default execution engine, integrating features like automatic KV cache sizing and batch-sharded sampling. While the release removes support for Medusa, MLP speculator, and the legacy Ray executor, it also delivers substantial performance gains through scheduler optimizations, achieving up to a 2.47x increase in output tokens per second without altering underlying kernels. AI

IMPACT Requires infrastructure updates for users, potentially improving serving efficiency and reducing latency for LLM deployments.

RANK_REASON This is a software release for an infrastructure component (vLLM) that impacts model serving, not a new frontier model release or core research.

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

vLLM 0.33.0 mandates CUDA 13, adopts Model Runner V2, and boosts speed

How we ranked this

Signal score
1 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
This is a software release for an infrastructure component (vLLM) that impacts model serving, not a new frontier model release or core research.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
infra, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
1 days old
Coverage has settled into its steady-state source set.

Full methodology in our editorial standards.

COVERAGE [1]

  1. dev.to — LLM tag TIER_1 English(EN) · Daniel Sam Pete Thiyagu ·

    vLLM 0.33, Plainly: CUDA 13, Model Runner V2, and What Breaks on Upgrade

    <p>vLLM 0.33.0 shipped on November 2, 2026 — 927 commits from 389 contributors — and the headline is not a model. It is a platform break: the wheels and Docker images are CUDA 13.0 only, PyTorch moves to 2.14.0, and three things your deployment may depend on are deleted outright.…

RELATED ENTITIES

RELATED TOPICS