Audio-model decision manipulation via unweighted downmixing
Published Apr 2, 2026 · Updated Apr 2, 2026
Improper input validation in vLLM 0.5.5 and later, before 0.18.0, allows remote authenticated users to manipulate audio-model decisions. The Librosa to_mono path averages every channel with numpy.mean instead of applying speaker-aware downmix weights. Crafted multichannel audio can sound benign in ordinary playback while its LFE or later channels alter transcription, moderation, or voice-authentication results.
Summary
What happened
Improper input validation in vLLM 0.5.5 and later, before 0.18.0, allows remote authenticated users to manipulate audio-model decisions. The Librosa to_mono path averages every channel with numpy.mean instead of applying speaker-aware downmix weights. Crafted multichannel audio can sound benign in ordinary playback while its LFE or later channels alter transcription, moderation, or voice-authentication results.
The record
- CVE
- CVE-2026-34760
- Published
- Apr 2, 2026
- Updated
- Apr 2, 2026
- Vendor
- vLLM Project
- Product
- vLLM
- Classifications
- CWE-20, T1027
- Attack vector
- network
- Privileges
- authenticated
Timeline
How it unfolded
- Apr 2, 2026CVE publishedPublication date reported by the CVE source.
- Apr 2, 2026Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=>= 0.5.5, < 0.18.0
What conditions does exploitation require?
What is affected?
Published CVSS scores
CVSS describes severity. EPSS estimates exploitation probability.
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Public exploit references
No public exploit references are available in this record.
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo