{"id":"CVE-2026-100653","title":"vLLM is an inference and serving engine for large language models","summary":"vLLM is an inference and serving engine for large language models. In versions from 0.22.1 through 0.28.0, the operator-supplied model revision pin (--revision / --code-revision) is not propagated to several Hugging Face artifact loads f…","severity":"medium","cvss":6.5,"cvssVector":"CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:H/A:N","cwe":["CWE-348","CWE-494"],"vendor":"vllm-project","product":"vllm","affected":["vllm >= 0.22.1 < 0.28.0"],"published":"2026-09-26","updated":"2026-09-30","sourceUpdated":"2026-09-30T18:18:01.040","source":"NVD","sourceUrl":"https://nvd.nist.gov/vuln/detail/CVE-2026-100653","references":[{"url":"https://github.com/vllm-project/vllm/commit/d26a28ab033697f55a1414b5b0435de7cd6045b6","label":"disclosure@vulncheck.com"},{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-hhv2-872h-628q","label":"disclosure@vulncheck.com"},{"url":"https://www.vulncheck.com/advisories/vllm-0.22.1-before-0.28.0-incomplete-artifact-pin-propagation","label":"disclosure@vulncheck.com"},{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-hhv2-872h-628q","label":"134c704f-9b21-4f2e-91b3-4a467353bcc0"},{"url":"https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-100653.json"},{"url":"https://access.redhat.com/security/cve/CVE-2026-100653"},{"url":"https://bugzilla.redhat.com/show_bug.cgi?id=2541818"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-100653"},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-100653"}],"tags":["nvd","cve.org","csaf","vex","red-hat","exploit-available","osv","pip"],"epss":0.0034,"epssPercentile":0.27254,"exploitAvailable":true,"ssvc":{"exploitation":"poc","automatable":"no","technicalImpact":"partial","timestamp":"2026-09-30T17:46:29.258369Z"},"ingestedAt":"2026-09-26T13:33:20.919Z","aliases":["PYSEC-2026-4324","GHSA-hhv2-872h-628q"],"ecosystem":"pip","patched":["vllm 0.28.0"],"slug":"CVE-2026-100653","body":"## Overview\n\nvLLM is an inference and serving engine for large language models. In versions from 0.22.1 through 0.28.0, the operator-supplied model revision pin (--revision / --code-revision) is not propagated to several Hugging Face artifact loads for the FunAudioChat and Tarsier2 architectures: the WhisperFeatureExtractor and speech_tokenizer PreTrainedTokenizerFast loads in vllm/model_executor/models/funaudiochat.py and the Qwen2VLConfig.from_pretrained call used by Tarsier2ProcessingInfo in vllm/model_executor/models/qwen2_vl.py. As a result, deployments pinned to a reviewed revision still resolve these behavior-affecting processor, tokenizer, and config artifacts from the repository's default revision, so a later change to the upstream default branch can alter audio preprocessing, speech tokenizer behavior, or Tarsier2 configuration without any change to the operator's configured pin. This is a supply-chain integrity and reproducibility failure for pinned deployments; it is residual to the earlier fix tracked as GHSA-3ww4-5jv9-j5gm / CVE-2026-47155 and does not constitute remote code execution or a trust_remote_code=False bypass. The issue is fixed in version 0.28.0.\n\n## Remediation\n\nRefer to the linked advisories for vendor-supplied fixes and affected version ranges.\n\n## Vendor advisories\n\n- **Red Hat VEX** · Moderate · affected: Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI) 3, Red Hat OpenShift AI (RHOAI) · no fix planned: Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI) 3, Red Hat OpenShift AI (RHOAI) · updated 2026-10-01 · [vex](https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-100653.json)\n\n## Package advisory (CVE-2026-100653)\n\nAffected packages:\n\n- `vllm >= 0.22.1, < 0.28.0`\n\nPatched in:\n\n- `vllm 0.28.0`\n\nSource: https://osv.dev/vulnerability/PYSEC-2026-4324","depth":"twilight","depthScore":48,"depthScoreParts":{"impact":35.8,"likelihood":0.1,"exploitation":12,"ransomware":0},"changes":[{"seq":213837,"id":"CVE-2026-100653","ts":1790792428518,"field":"exploit_available","old":"false","new":"true"}]}