CSIRTS // UNIFIED SECURITY ADVISORY FEEDSYS ● ONLINE · POWERED BY INTELFUSIONS.COM

CVE-2026-55646: vLLM is an inference and serving engine for large language models. From 0.22.0 to 0.23.0, the /v1/audio/transcriptions and /v1/audio/translations routes call request.file.read() to

mediumCVSS 6.5CVE-2026-55646
vLLM is an inference and serving engine for large language models. From 0.22.0 to 0.23.0, the /v1/audio/transcriptions and /v1/audio/translations routes call request.file.read() to fully materialize an uploaded audio file into memory before vLLM checks the documented VLLM_MAX_AUDIO_CLIP_FILESIZE_MB compressed upload size limit (default 25 MB) later in the speech-to-text preprocessing step, so an API caller who can reach those routes can submit an oversized multipart upload and cause vLLM to allocate memory proportional to the uploaded file size before the request is rejected as too large, creating memory pressure or terminating the process depending on deployment resource limits. This issue is fixed in version 0.24.0.

Details

Source
NVD Recent CVEs (US · database · site)
Severity
medium — CVSS 6.5
Published
2026-07-06
Last updated
2026-07-07
Exploitation
Not in CISA KEV at last sync

Original advisory: https://nvd.nist.gov/vuln/detail/CVE-2026-55646

Exploitation outlook

EPSS (FIRST.org) estimates each CVE’s probability of exploitation in the next 30 days — here is the CSIRTS.com read on those numbers.

Referenced CVEs

CVECSIRTS overviewExternal
CVE-2026-55646coverage & exploitation statusNVD · CVE.org

Same CVEs, other sources

How other CERTs, PSIRTs and databases cover the vulnerabilities in this advisory.

Recent advisories for vLLM is an

A cluster of recent advisories against the same product widens the attack surface — attackers routinely chain freshly published CVEs on one product, so review these together.

More from NVD Recent CVEs