CVE-2026-116178: vLLM 0.20.0 before 0.30.0 DoS via /inference/v1/generate mm_placeholders

Q: Which deployments are exposed to this issue?

vLLM versions from 0.20.0 up to, but not including, 0.30.0 are affected when the disaggregated /inference/v1/generate endpoint is reachable by a remote attacker. The vulnerable processing path involves multimodal metadata and Qwen2.5-VL position construction.

Q: What does an attacker need to send to trigger the denial of service?

An unauthenticated remote attacker can submit a valid multimodal item with inconsistent metadata. Examples include repeated or overlapping mm_placeholders ranges, or mismatched kwargs_data and mm_hashes lists, which can create a negative text span and terminate the engine.

Q: What is the operational impact of successful exploitation?

Successful exploitation crashes or terminates the vLLM engine, causing a denial of service. The provided information does not indicate confidentiality or integrity impact.

Published Oct 25, 2026

Updated

vLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that allows remote attackers to crash the engine by sending inconsistent multimodal metadata to the disaggregated /inference/v1/generate endpoint. Attackers can replay a valid multimodal item with repeated or overlapping mmplaceholders ranges or mismatched kwargsdata and mmhashes lists, producing a negative text span in Qwen2.5-VL position construction that terminates the engine.

Affected Software

1 affected component

vllm vllm>=0.20.0<0.30.0

Event History

Oct 25, 2026

CVE Published

via MITRE·01:20 PM

Data Sourced

via MITRE·01:20 PM

DescriptionSeverityWeakness

Frequently Asked Questions

Which deployments are exposed to this issue?

vLLM versions from 0.20.0 up to, but not including, 0.30.0 are affected when the disaggregated /inference/v1/generate endpoint is reachable by a remote attacker. The vulnerable processing path involves multimodal metadata and Qwen2.5-VL position construction.

What does an attacker need to send to trigger the denial of service?

An unauthenticated remote attacker can submit a valid multimodal item with inconsistent metadata. Examples include repeated or overlapping mm_placeholders ranges, or mismatched kwargs_data and mm_hashes lists, which can create a negative text span and terminate the engine.

What is the operational impact of successful exploitation?

Successful exploitation crashes or terminates the vLLM engine, causing a denial of service. The provided information does not indicate confidentiality or integrity impact.

CVE-2026-116178: vLLM 0.20.0 before 0.30.0 DoS via /inference/v1/generate mm_placeholders

Affected Software

Event History

Frequently Asked Questions

Which deployments are exposed to this issue?

What does an attacker need to send to trigger the denial of service?

What is the operational impact of successful exploitation?

Company

Resources

Contact

Frequently Asked Questions

Which deployments are exposed to this issue?

What does an attacker need to send to trigger the denial of service?

What is the operational impact of successful exploitation?