HighVulnerabilityLLM-specific
CVE-2026-130685: vLLM denial of service through out-of-range prompt token ids in /v1/completions
- Source
- NVD(opens in a new tab)
- Identifier
- CVE-2026-130685
- Published
- Record updated
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
Summary
vLLM from 0.28.0 before 0.33.0 contains an improper input validation flaw. Unauthenticated attackers can send an out-of-range prompt token id, such as a large positive value, to /v1/completions, /v1/embeddings, or /pooling when the server runs with --skip-tokenizer-init, which bypasses the tokenizer-gated vocab check. The resulting fatal embedding index error crashes the shared EngineCore and denies service to all tenants.
Mitigation
The source does not state a fix yet. Check the original advisory for updates.
Topics
Related items
- HighCVE-2026-130686: vLLM resource exhaustion through many remote image URLs in one requestSimilar attack · NVD/CVE Database
- MediumCVE-2026-126399: vLLM resource exhaustion via video frame limit bypass in chat completionsSimilar attack · NVD/CVE Database
- MediumCVE-2026-126398: vLLM reachable assertion crash in GPUModelRunner via embeddings-only requestsSimilar attack · NVD/CVE Database
- HighCVE-2026-126397: vLLM denial of service via unbounded max_patches in multimodal preprocessingSimilar attack · NVD/CVE Database
- MediumCVE-2026-126396: vLLM resource exhaustion in MiniMax M3 VL video processor via min_pixelsSimilar attack · NVD/CVE Database