---
id: "CVE-2026-126398"
page: "https://infrastack.ai/live/CVE-2026-126398"
kind: "vulnerability"
aliases: ["CVE-2026-126398"]
title: "vLLM: embeddings-only request crashes the engine for all clients"
severity: "high"
cvss: 8.2
epss: null
exploited: false
status: "open"
category: "ai-tooling"
products: []
ecosystems: []
tags: ["ai-tooling", "llm-serving", "python", "self-hosted", "vllm"]
not_affected_if: ["default-config", "non-xdrope-models"]
published_at: "2026-11-13T10:17:44+00:00"
first_seen_at: "2026-11-13T10:47:19+00:00"
updated_at: "2026-11-13T10:52:19+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-126398", "first_seen_at": "2026-11-13T10:47:19+00:00"}]
exploit_sources: []
---

# vLLM: embeddings-only request crashes the engine for all clients

Link for people: https://infrastack.ai/live/CVE-2026-126398 (this .md is for agents)

vLLM from 0.28.0 before 0.30.0 has a reachable assertion in GPUModelRunner._init_xdrope_positions. A remote attacker who sends a /v1/completions request containing only prompt_embeds can crash the EngineCore and API server, cutting off all clients.

_Summary, check and fix written by a model from the sources; labelled as such._

## How to check

- Check the installed vLLM version with `pip show vllm` and whether it falls in 0.28.0 up to but not including `0.30.0`.
- Check whether the server runs with `--enable-prompt-embeds` and serves XD-RoPE models.

## Fix

- vLLM: move to `0.30.0` or later (the sources state the fixed version as 0.30.0).

## Timeline

- 2026-11-13T10:47:19+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-126398

## Documents

- [advisory] Maintainer's advisory: github.com https://github.com/vllm-project/vllm/security/advisories/GHSA-5q78-j82c-2vr2
- [advisory] Third-party advisory: www.vulncheck.com https://www.vulncheck.com/advisories/vllm-0.28.0-before-0.30.0-denial-of-service-via-prompt-embeds-on-xd-rope-models

## From the sources

vLLM from 0.28.0 before 0.30.0 contains a reachable assertion vulnerability in GPUModelRunner._init_xdrope_positions that allows remote attackers to crash the engine by submitting embeddings-only requests. Attackers can send a /v1/completions request containing only prompt_embeds to XD-RoPE models with --enable-prompt-embeds, terminating the EngineCore and API server for all clients.

---

_Free to read, with fair use of about 60 pages a minute. For higher limits or API access: sales@infrastack.ai_