vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing (to_mono), while the international standard ITU-R BS.775-4 specifies a weighted downmixing algorithm. This discrepancy results in inconsistency between audio heard by humans (e.g., through headphones/regular speakers) and audio processed by AI models (Which infra via Librosa, such as vllm, transformer). This issue has been patched in version 0.18.0.
Advisories

No advisories yet.

Fixes

Solution

No solution given by the vendor.


Workaround

No workaround given by the vendor.

History

Thu, 02 Apr 2026 20:30:00 +0000

Type Values Removed Values Added
Description vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing (to_mono), while the international standard ITU-R BS.775-4 specifies a weighted downmixing algorithm. This discrepancy results in inconsistency between audio heard by humans (e.g., through headphones/regular speakers) and audio processed by AI models (Which infra via Librosa, such as vllm, transformer). This issue has been patched in version 0.18.0.
Title vLLM: Downmix Implementation Differences as Attack Vectors Against Audio AI Models
Weaknesses CWE-20
References
Metrics cvssV3_1

{'score': 5.9, 'vector': 'CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:H/A:L'}


Projects

Sign in to view the affected projects.

cve-icon MITRE

Status: PUBLISHED

Assigner: GitHub_M

Published:

Updated: 2026-04-02T18:59:49.638Z

Reserved: 2026-03-30T19:17:10.225Z

Link: CVE-2026-34760

cve-icon Vulnrichment

No data.

cve-icon NVD

Status : Received

Published: 2026-04-02T20:16:25.437

Modified: 2026-04-02T20:16:25.437

Link: CVE-2026-34760

cve-icon Redhat

No data.

cve-icon OpenCVE Enrichment

No data.

Weaknesses