CVE-2025-62372: vLLM engine crash via malformed multimodal embedding input shape

Summary

CVE-2025-62372 affects vLLM, an inference and serving engine for large language models, from version 0.5.5 before 0.11.1. Passing multimodal embedding inputs with the correct ndim but an incorrect shape, such as a wrong hidden dimension, can crash the vLLM engine serving multimodal models. The flaw is classified as CWE-129, Improper Validation of Array Index, and GitHub rates it CVSS 4.0 8.3 (High), with availability impact only.

Mitigation

This issue has been patched in version 0.11.1.

Affected packages in the Exposure Registry

Matched by package name and ecosystem. Each entry shows whether the package delegates to a language model and how many tracked packages depend on it.

• vllmPyPILLM dependency since 2023-08-25 · 15 tracked dependents

Topics

• Inference infrastructure

Related items

• LowCloakFEM: Agentic FEM Code Generation for Red-Teaming Computational Resource ExhaustionSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
• MediumThe Rise of LLMJacking: How Threat Actors Monetize Compromised Cloud and Model ResourcesSimilar attack · Google Threat Intelligence
• InfoThe Blind Spots of Detection: Adversarial Perturbation Attacks against Watchdogs in Distributed LLM InferenceSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
• HighCVE-2026-118761: Kubeflow Pipelines decompression bomb in pipeline archive uploadSimilar attack · NVD/CVE Database
• LowOpenCode User Hit with Hostile Backlash After Reporting Free AI LoopholeSimilar attack · Socket Blog