CVE-2026-71486Public exploitNetwork attack vectorHinoki check available

Service interrupt via unchecked derender decoding

Published Aug 17, 2026 · Updated Aug 18, 2026

Denial of service in vLLM before 0.26.0 allows remote authenticated users to exhaust CPU and memory via derender API requests. The /v1/completions/derender and /v1/chat/completions/derender handlers pass caller-supplied GenerateResponse choices, token_ids, and nested logprob data into OnlineDerenderer and tokenizer.decode before enforcing max_model_len, max_tokens, max_num_seqs, choice-count, or response-size bounds. Exposure requires access to the OpenAI-compatible /v1 derender routes, typically behind the API-key middleware when configured, and the documented consequence is oversized responses and availability degradation rather than code execution or data disclosure.

CVSS severity4.3
Medium
EPSS probability0.34%
Next 30 days · Sep 16, 2026
Known exploitationUnconfirmed
Based on sourced intelligence
Hinoki checkAvailable
Coverage for this vulnerability

See if you're affected

Explore vulnerability checks for your environment with Hinoki.

Book a demo

Summary

What happened

Denial of service in vLLM before 0.26.0 allows remote authenticated users to exhaust CPU and memory via derender API requests. The /v1/completions/derender and /v1/chat/completions/derender handlers pass caller-supplied GenerateResponse choices, token_ids, and nested logprob data into OnlineDerenderer and tokenizer.decode before enforcing max_model_len, max_tokens, max_num_seqs, choice-count, or response-size bounds. Exposure requires access to the OpenAI-compatible /v1 derender routes, typically behind the API-key middleware when configured, and the documented consequence is oversized responses and availability degradation rather than code execution or data disclosure.

The record

CVE
CVE-2026-71486
Published
Aug 17, 2026
Updated
Aug 18, 2026
Vendor
vLLM Project
Product
vLLM
Classifications
CWE-400, CWE-770, T1499
Attack vector
network
Privileges
authenticated

Timeline

How it unfolded

  1. Aug 17, 2026CVE publishedPublication date reported by the CVE source.
  2. Aug 18, 2026Record updatedLatest update available in the CVE record.

Exploitability

Present is not the same as exploitable

Compare your product and version with the public record. A matching version still requires validation against your environment.

Is a vulnerable build present?

Compare these published version ranges with your installed build and any vendor patches.

  1. Affected versionversion=< 0.26.0

What conditions does exploitation require?

Attack vectornetwork
Required privilegesauthenticated

What is affected?

vLLM Project · vLLMversion=< 0.26.0

Published CVSS scores

4.3GitHub, Inc.CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L

CVSS describes severity. EPSS estimates exploitation probability.

Attacks

What attackers are doing with it

Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.

Daily unique IPsNo honeypot observations are available for this CVE in the selected window.

No observations available

Sep 9, 2026Sep 15, 2026
Latest reporting daySep 15, 2026
Latest daily unique IPsUnavailable
Prior 30-day averageUnavailable
SourceShadowserver honeypots (KEV)
Vectornetwork
Privilegesauthenticated
Known exploitationUnconfirmed
Public exploitPublished

Weakness, pattern, technique

CWE-400Uncontrolled Resource Consumption
CWE-770Allocation of Resources Without Limits or Throttling
T1499Endpoint Denial of Service

Public exploit references

Labels summarize the accepted research assessment. They do not indicate a test against your environment.

Technologies

Your stack

See the directory against your own environment.

Your stack

Check the software in your environment

Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.

Book a demo