PatchSiren cyber security CVE debrief
CVE-2026-116173 vllm-project CVE debrief
A denial of service vulnerability exists in vLLM versions from 0.22.0 before 0.24.0. Remote attackers can terminate EngineCore by bypassing logprob_token_ids limits in the Rust gRPC frontend. This is achieved by sending Generate or GenerateStream requests with output_logprobs enabled and more than 128 output_candidates.token_ids, triggering a fatal sampler exception that crashes the engine.
- Vendor
- vllm-project
- Product
- vllm
- CVSS
- HIGH 8.2
- CISA KEV
- Not listed in stored evidence
- Original CVE published
- 2026-10-25
- Original CVE updated
- 2026-10-29
- Advisory published
- 2026-10-25
- Advisory updated
- 2026-10-29
Who should care
Defenders responsible for systems using vLLM versions from 0.22.0 before 0.24.0 should assess exposure and prioritize remediation or compensating controls. This includes operators, platform administrators, vulnerability management teams, and security teams who need to verify vLLM versions, review system configurations, and apply immediate remediation or compensating controls to prevent potential denial of service attacks.
Why it matters
Defenders should care about CVE-2026-116173 because it affects vLLM versions from 0.22.0 before 0.24.0, allowing remote attackers to terminate EngineCore. This requires verification of vLLM versions, assessment of exposure, and immediate remediation or compensating controls to prevent potential denial of service.
- Potential denial of service due to engine crashes.
- Need for verification of vLLM version and exposure to untrusted input.
- Potential impact on system availability and reliability.
- Requires immediate remediation or compensating controls for vulnerable versions.
Technical summary
The vulnerability exists in the Rust gRPC frontend of vLLM, allowing remote attackers to terminate EngineCore by bypassing logprob_token_ids limits. This is achieved by sending specific requests that trigger a fatal sampler exception, crashing the engine. Affected versions include vLLM from 0.22.0 before 0.24.0. Defenders should prioritize verifying the version of vLLM in use and assessing exposure to untrusted input to prevent potential denial of service attacks through engine crashes and impact on system availability and reliability.
Defensive priority
Defenders should prioritize verifying the version of vLLM in use and assessing exposure to untrusted input. Immediate remediation or compensating controls should be applied to systems using vulnerable versions.
Recommended defensive actions
- Verify the version of vLLM in use and assess exposure to untrusted input.
- Apply immediate remediation or compensating controls to systems using vulnerable versions.
- Monitor for suspicious Generate or GenerateStream requests with output_logprobs enabled.
- Review system configurations for potential exposure.
- Check relevant monitoring, detection, and logs for exposed assets that need extra review.
- Track exceptions, retest remediated assets, and close the item only after evidence is documented.
- Confirm whether affected product deployments exist in managed environments and assign an owner for follow-up.
Evidence notes
The vulnerability details are based on the source item from osv_dev and supplemental sources. The CVE Program and NVD records provide additional context. Evidence limits suggest verifying vLLM version 0.22.0 to 0.24.0 for potential exposure. Defenders should review official advisories and assess system configurations for Generate or GenerateStream requests with output_logprobs enabled.
Sources and references
Verified primary and authoritative sources
-
CVE-2026-116173 CVE Program record
Publisher, destination, and source semantics verified
URL: https://www.cve.org/CVERecord?id=CVE-2026-116173
CVE Program - Official CVE Program record with source-provided CVE metadata.
-
CVE-2026-116173 NVD vulnerability detail
Publisher, destination, and source semantics verified
URL: https://nvd.nist.gov/vuln/detail/CVE-2026-116173
NIST National Vulnerability Database - Official NIST NVD detail page and source-specific vulnerability assessment.
Supplemental references
-
PYSEC-2026-4350
Unverified legacy reference
URL: https://storage.googleapis.com/osv-vulnerabilities/PyPI/PYSEC-2026-4350.json
osv_dev
-
Source reference
Unverified legacy reference
URL: https://github.com/vllm-project/vllm/security/advisories/GHSA-6pgw-f3pv-4h8q
Supplemental source
-
Source reference
Unverified legacy reference
URL: https://www.vulncheck.com/advisories/vllm-0.22.0-before-0.24.0-denial-of-service-via-grpc-output-candidates-token-ids
Supplemental source
Methodology and review provenance
AI-assisted synthesis based on stored public vulnerability evidence. System validation, approval state, and publication status do not by themselves establish human review of this revision. PatchSiren helps prioritize defensive review and does not prove exposure or remediation on any system.