vLLM Stop Token ID Boundary Validation Flaw Causes Service Crash
vLLM
vLLM versions 0.22.0 through 0.23.0 fail to validate stop_token_ids against vocabulary bounds. Attackers can submit crafted requests that trigger CUDA tensor indexing failures and crash the inference engine.
disclosed CVENo confirmed exploitation, so this record is tracked as a disclosed vulnerability and not counted as an incident. It moves into the counts if CISA or the vendor confirms exploitation. Why
- CVSS
- 8.2
- EPSS
- <0.01 · top 64% · 2026-11-13
- fix status
- Fix status unknown
- first report
- 26 Sep 2026
- last update
- 26 Sep 2026
- CVE ids
Timeline · 1
Sources · 1
Advisories and databases
- 2026-09-26CVE-2026-100652: vLLM versions 0.22.0 through 0.23.0 fail to validate stop_token_ids against vocabulary bounds in Rust HTTP and gRPC frontends,…NVD · nvd.nist.gov/vuln/detail/CVE-2026-100652
Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.
Classification
- lane
- Patch now
- AI role
- Attack on AI
- categories
- Infrastructure vulnerability
- severity
- 2 · Disclosed vulnerability
- fix status
- Fix status unknown
- targets
- Framework / tooling
- owasp llm top 10
- LLM03:2025 Supply Chain
- mitre atlas
- no technique for these categories
- dates
- happened 26 Sep 2026 · first reported 26 Sep 2026
- origin
- NVDVulnerability record from NVD / the GitHub Advisory Database. This product uses the NVD API but is not endorsed or certified by the NVD.
- record
- confidence 0.95 · created 2026-10-18 06:33Z · updated 2026-10-18 06:33Z
Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned
Copy a brief
This record as a short Markdown brief, for a ticket, a chat or a coding agent.
the brief as text
# vLLM Stop Token ID Boundary Validation Flaw Causes Service Crash https://gcdradar.vercel.app/incidents/vllm-stop-token-id-boundary-validation-flaw-causes-service-c · AIR-2026-0211 · closed 2026-09-26 - Lane: Patch now - Severity: 2 (Disclosed vulnerability) - Fix status: Fix status unknown - First reported: 2026-09-26 - AI role: Attack on AI - Products: vLLM - CVE ids: CVE-2026-100652 - CVSS: 8.2 ## Summary vLLM versions 0.22.0 through 0.23.0 fail to validate stop_token_ids against vocabulary bounds. Attackers can submit crafted requests that trigger CUDA tensor indexing failures and crash the inference engine. ## What to do No action recorded; read the sources below and the vendor's advisory. ## Latest updates - 2026-09-26: CVE-2026-100652: vLLM versions 0.22.0 through 0.23.0 fail to validate stop_token_ids against vocabulary bounds in Rust HTTP and gRPC frontends,… (https://nvd.nist.gov/vuln/detail/CVE-2026-100652) ## Sources - 2026-09-26 NVD: CVE-2026-100652: vLLM versions 0.22.0 through 0.23.0 fail to validate stop_token_ids against vocabulary bounds in Rust HTTP and gRPC frontends,… https://nvd.nist.gov/vuln/detail/CVE-2026-100652 Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.
Related
- closed7.5vLLM CVE-2026-115912: vLLM shared-memory multimodal remote code execution
Patch now · Disclosed · Patched · since 26 Oct 2026 · updated 28 Oct 2026 · CVE-2026-115912
AI - closed6.5vLLM CVE-2026-115916: vLLM tokenizer off-by-one engine termination
Patch now · Disclosed · Patched · since 26 Oct 2026 · updated 28 Oct 2026 · CVE-2026-115916
AI - closed6.5vLLM CVE-2026-115918: vLLM chat template parsing resource exhaustion
Patch now · Disclosed · Patched · since 26 Oct 2026 · updated 28 Oct 2026 · CVE-2026-115918
AI - closed6.5vLLM CVE-2026-115914: vLLM multimodal OCR path denial of service
Patch now · Disclosed · Patched · since 26 Oct 2026 · updated 28 Oct 2026 · CVE-2026-115914
AI - closed8.1vLLM CVE-2026-115915: vLLM insecure image parsing command execution vulnerability
Patch now · Disclosed · Patched · since 26 Oct 2026 · updated 28 Oct 2026 · CVE-2026-115915
AI - closed3.4vLLM CVE-2026-115917: vLLM permissive CORS configuration vulnerability
Patch now · Disclosed · Patched · since 26 Oct 2026 · updated 28 Oct 2026 · CVE-2026-115917
AI