---
id: "CVE-2026-116173"
kind: "vulnerability"
aliases: ["CVE-2026-116173"]
title: "vLLM: oversized logprob requests over gRPC crash the inference engine"
severity: "high"
cvss: 8.2
epss: null
exploited: false
status: "open"
category: "ai-tooling"
products: []
ecosystems: []
tags: ["ai-tooling", "grpc", "inference", "python", "self-hosted", "vllm"]
not_affected_if: ["default-config", "internal-only"]
published_at: "2026-10-25T14:16:56+00:00"
first_seen_at: "2026-10-25T14:41:23+00:00"
updated_at: "2026-10-25T14:42:26+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-116173", "first_seen_at": "2026-10-25T14:41:23+00:00"}]
---

# vLLM: oversized logprob requests over gRPC crash the inference engine

vLLM from 0.22.0 before 0.24.0 has a denial of service flaw in its Rust gRPC frontend. A remote request with output logprobs enabled and more than 128 candidate token IDs makes the sampler throw a fatal exception and terminates EngineCore.

_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 see whether it falls in 0.22.0 up to but not including 0.24.0.
- Check whether the gRPC frontend is exposed: look for the Generate and GenerateStream endpoints reachable from outside the host.

## Fix

- Upgrade vLLM to `0.24.0` or later.
- Until upgraded, block external access to the gRPC Generate and GenerateStream endpoints.

## Timeline

- 2026-10-25T14:41:23+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-116173

## From the sources

vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote attackers to terminate EngineCore by bypassing logprob_token_ids limits in the Rust gRPC frontend. Attackers can send Generate or GenerateStream requests with output_logprobs enabled and more than 128 output_candidates.token_ids, triggering a fatal sampler exception that crashes the engine.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_