vLLM denial of service via out-of-vocabulary stop token IDs
vllm-project vllm
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| vllm-project vllm |
|
|
Original description
vLLM from 0.20.1 before 0.29.0 contains an improper input validation vulnerability that allows remote attackers to crash the engine by supplying out-of-vocabulary stop_token_ids. Attackers can send OpenAI chat, completion, generate, or batch requests with invalid IDs to terminate the engine or write outside the V2 runner's logits buffer.
From the CVE record · VulnCheck
References
- github.com/vllm-project/vllm/security/advisories/GHSA-w9r6-29q4-pvprVendor advisory
- vulncheck.com/advisories/vllm-0.20.1-before-0.29.0-dos-and-gpu-out-of-bounds-write-via-stop-token-idsAdvisory
- github.com/CVEProject/cvelistV5/tree/main/cves/2026/116xxx/CVE-2026-116181.json
- nvd.nist.gov/vuln/detail/CVE-2026-116181