# CVE-2026-114055: vLLM uncontrolled resource consumption in multimodal processor

Source: https://junglewise.ai/threats/cve-2026-114055-vllm-from-0-7-3-before-0-31-0-contains-an-uncontrolled-resource  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-10-21. Updated: 2026-10-21  
License: CC BY 4.0 with a link back

CVE-2026-114055 · Severity: high · CVSS 7.5 · Published 2026-10-21

vLLM is an inference engine for large language models that supports multimodal inputs like images and videos. A vulnerability allows remote attackers to supply malicious image or video processing parameters that force excessive memory allocation, causing the shared server to crash and deny service to all users. No authentication is required to exploit this issue.

Technologies: [vllm (PyPI)](https://junglewise.ai/threats/technologies/pypi-vllm), [vLLM Project vllm](https://junglewise.ai/threats/technologies/vllm-project-vllm). Vendors: [PyPI](https://junglewise.ai/threats/vendors/pypi), [vLLM Project](https://junglewise.ai/threats/vendors/vllm-project).

## Quick answers

**Is CVE-2026-114055 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-114055 being exploited in the wild.

**How severe is CVE-2026-114055?** High severity, with a CVSS score of 7.5, published 21 October 2026.

**Which products and versions are affected?** vllm (PyPI) >= 0.7.3, < 0.31.0, ECOSYSTEM:introduced=0.7.3,fixed=0.31.0; vLLM Project vllm 0.7.3 before 0.31.0.

**Is there a fix for CVE-2026-114055?** Yes. vllm (PyPI): upgrade to 0.31.0 or later. Until then: Restrict access to the vLLM API server with an external authentication layer if it is exposed to untrusted clients. Alternatively, disable per-request mm_processor_kwargs by default and only enable the override behavior for deployments serving only trusted clients. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-114055 attacked?** It is attacked over the network.

**Am I affected by CVE-2026-114055?** Check whether you run vllm (PyPI) or vLLM Project vllm at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## How to fix

- vllm (PyPI): upgrade to 0.31.0 or later

  ```sh
  pip install "vllm>=0.31.0"
  ```


Restrict access to the vLLM API server with an external authentication layer if it is exposed to untrusted clients. Alternatively, disable per-request mm_processor_kwargs by default and only enable the override behavior for deployments serving only trusted clients.

## Technical details

The vulnerability exists in vLLM's handling of per-request mm_processor_kwargs parameters, which are forwarded without validation to multimodal image and video processors (Isaac, InternVL, H2OVL, and others). An unauthenticated remote attacker can set large values for image size, patch count, crop count, or frame count to trigger out-of-memory conditions and terminate the service process. The vulnerability affects all multimodal processor implementations due to unvalidated constructor argument forwarding at the shared API boundary.

## Affected products

- vLLM Project vLLM 0.7.3 before 0.31.0

## Identifiers

- CVE-2026-114055
- GHSA-gx7p-2j49-hfq4

## Timeline

- 2026-10-06: disclosed
- 2026-10-21: advisory

## References

- [https://github.com/vllm-project/vllm/security/advisories/GHSA-gx7p-2j49-hfq4](https://github.com/vllm-project/vllm/security/advisories/GHSA-gx7p-2j49-hfq4)
- [https://www.vulncheck.com/advisories/vllm-0.7.3-before-0.31.0-dos-via-unvalidated-mm-processor-kwargs](https://www.vulncheck.com/advisories/vllm-0.7.3-before-0.31.0-dos-via-unvalidated-mm-processor-kwargs)

## Related threats

- [vLLM chat template resource exhaustion DoS](https://junglewise.ai/threats/cve-2026-115918-vllm-is-an-inference-and-serving-engine-for-large-language)
- [vLLM: Permissive default CORS origins allow cross-origin requests to the vLLM HTTP server](https://junglewise.ai/threats/cve-2026-115917-vllm-is-an-inference-and-serving-engine-for-large-language)
- [vLLM off-by-one error in tokenizer max_token_id validation](https://junglewise.ai/threats/cve-2026-115916-vllm-is-an-inference-and-serving-engine-for-large-language)
- [vLLM image format validation bypass in media processing](https://junglewise.ai/threats/cve-2026-115915-vllm-is-an-inference-and-serving-engine-for-large-language)
- [vLLM denial of service in DeepSeek-OCR image processing](https://junglewise.ai/threats/cve-2026-115914-vllm-is-an-inference-and-serving-engine-for-large-language)