CVE-2026-71486: vLLM derender endpoints vulnerable to resource exhaustion

# CVE-2026-71486: vLLM derender endpoints vulnerable to resource exhaustion

https://gcdradar.vercel.app/incidents/cve-2026-71486-vllm-derender-endpoints-vulnerable-to-resourc · AIR-2026-0083 · closed 2026-08-17

- Lane: Patch now
- Severity: 2 (Disclosed vulnerability)
- Fix status: Fix status unknown
- First reported: 2026-08-17
- AI role: Attack on AI
- Products: vLLM
- CVE ids: CVE-2026-71486
- CVSS: 4.3

## Summary

The vLLM inference engine prior to version 0.26.0 contains a vulnerability in its derender endpoints that allows authenticated API clients to consume excessive CPU and memory. The issue has been addressed in version 0.26.0.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2026-09-04: CVE-2026-71486: vLLM: Derender endpoints decode caller-supplied GenerateResponse token IDs without output bounds (https://github.com/advisories/GHSA-8737-qx52-hjff)
- 2026-08-17: CVE-2026-71486: vLLM is an inference and serving engine for large language models. (https://nvd.nist.gov/vuln/detail/CVE-2026-71486)

## Sources

- 2026-08-17 NVD: CVE-2026-71486: vLLM is an inference and serving engine for large language models. https://nvd.nist.gov/vuln/detail/CVE-2026-71486
- 2026-09-04 GitHub Advisory Database: CVE-2026-71486: vLLM: Derender endpoints decode caller-supplied GenerateResponse token IDs without output bounds https://github.com/advisories/GHSA-8737-qx52-hjff

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.