# CVE-2026-115918: vLLM chat template resource exhaustion DoS

Source: https://junglewise.ai/threats/cve-2026-115918-vllm-is-an-inference-and-serving-engine-for-large-language  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-10-26. Updated: 2026-10-26  
License: CC BY 4.0 with a link back

CVE-2026-115918 · Severity: medium · CVSS 6.5 · Published 2026-10-26

vLLM is an inference engine for large language models that serves AI models over HTTP APIs. An authenticated attacker can cause the service to become unresponsive by injecting malicious template code into chat requests, exhausting CPU and memory resources. This allows a legitimate user with API access to disrupt service availability for all other users.

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-2026-115918 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-115918 being exploited in the wild.

**How severe is CVE-2026-115918?** Medium severity, with a CVSS score of 6.5, published 26 October 2026.

**Which products and versions are affected?** vllm (PyPI) >= 0.16.0, < 0.32.0, ECOSYSTEM:introduced=0.16.0,fixed=0.32.0; vLLM Project vllm 0.16.0 to 0.31.x.

**Is there a fix for CVE-2026-115918?** Yes. vllm (PyPI): upgrade to 0.32.0 or later. Until then: Set a server-side chat template configuration to prevent user template injection. Configure the VLLM_CHAT_TEMPLATE_RENDER_TIMEOUT environment variable to enforce a timeout on template rendering operations. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-115918 attacked?** It is attacked over the network, in a software library.

**Am I affected by CVE-2026-115918?** 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 0.32.0 or later

  ```sh
  pip install "vllm>=0.32.0"
  ```


Set a server-side chat template configuration to prevent user template injection. Configure the VLLM_CHAT_TEMPLATE_RENDER_TIMEOUT environment variable to enforce a timeout on template rendering operations.

## Technical details

An authentication bypass in the /v1/responses endpoint allows authenticated clients to bypass the trust_request_chat_template policy and inject attacker-controlled Jinja2 templates via chat_template_kwargs[chat_template]. The ChatParams.get_apply_chat_template_kwargs() function fails to enforce the intended trust policy, permitting unbounded Jinja rendering that causes CPU or memory exhaustion. The vulnerability affects versions 0.16.0 through 0.31.x when no server template is configured.

## Affected products

- vLLM Project vLLM 0.16.0 to 0.31.x

## Identifiers

- CVE-2026-115918
- GHSA-x567-49g5-p4cf
- PYSEC-2026-4349

## Timeline

- 2026-10-26: disclosed
- 2026-09-30: patched

## References

- [https://github.com/vllm-project/vllm/commit/cff08b461e086523cee8bba5bd3c66d302625134](https://github.com/vllm-project/vllm/commit/cff08b461e086523cee8bba5bd3c66d302625134)
- [https://github.com/vllm-project/vllm/pull/50300](https://github.com/vllm-project/vllm/pull/50300)

## Related threats

- [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)
- [vLLM Phi-4 multimodal audio encoder GPU memory exhaustion](https://junglewise.ai/threats/cve-2026-115913-vllm-is-an-inference-and-serving-engine-for-large-language)