Is CVE-2026-105757 exploited in the wild?

No. Junglewise Threat Intelligence has no report of CVE-2026-105757 being exploited in the wild.

How severe is CVE-2026-105757?

Medium severity, with a CVSS score of 6.5, published 5 October 2026.

Which products and versions are affected?

vllm (PyPI) < 0.30.0, ECOSYSTEM:introduced=0,fixed=0.30.0; Vllm prior to 0.30.0.

Is there a fix for CVE-2026-105757?

Yes. vllm (PyPI): upgrade to 0.30.0 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

How is CVE-2026-105757 attacked?

It is attacked over the network, in a software library.

Am I affected by CVE-2026-105757?

Check whether you run vllm (PyPI) or 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.

CVE-2026-105757: vLLM structured-output request handling denial of service

Q: Which products and versions are affected?

vllm (PyPI) < 0.30.0, ECOSYSTEM:introduced=0,fixed=0.30.0; Vllm prior to 0.30.0.

Q: Is there a fix for CVE-2026-105757?

Yes. vllm (PyPI): upgrade to 0.30.0 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

Q: How is CVE-2026-105757 attacked?

It is attacked over the network, in a software library.

Q: Am I affected by CVE-2026-105757?

Check whether you run vllm (PyPI) or 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.

CVE-2026-105757 · Severity: medium · CVSS 6.5 · Published 2026-10-05

vLLM is an inference engine for large language models that enables applications to generate text responses. A flaw in handling constrained text generation requests can allow attackers to submit malformed requests that crash the shared engine, causing a denial of service that affects all users relying on that instance.

Technologies: vllm (PyPI), Vllm. Vendors: PyPI, Vllm.

Quick answers

Is CVE-2026-105757 exploited in the wild?: No. Junglewise Threat Intelligence has no report of CVE-2026-105757 being exploited in the wild.
How severe is CVE-2026-105757?: Medium severity, with a CVSS score of 6.5, published 5 October 2026.
Which products and versions are affected?: vllm (PyPI) < 0.30.0, ECOSYSTEM:introduced=0,fixed=0.30.0; Vllm prior to 0.30.0.
Is there a fix for CVE-2026-105757?: Yes. vllm (PyPI): upgrade to 0.30.0 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
How is CVE-2026-105757 attacked?: It is attacked over the network, in a software library.
Am I affected by CVE-2026-105757?: Check whether you run vllm (PyPI) or 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 0.30.0 or later
```
pip install "vllm>=0.30.0"
```

Technical details

The vulnerability exists in vLLM's structured-output request validation prior to version 0.30.0, where multiple code paths can bypass request-scoped validation and reach the EngineCore fatal-error handler: backend mismatches trigger unhandled grammar compilation exceptions, the ngram_gpu speculative-decoding mode can pass invalid token indices to validation, and the Rust frontend accepts empty structured outputs that the Python frontend rejects. These conditions allow an attacker with network access to send specially crafted constrained-generation requests that terminate the shared engine.

Affected products

vLLM vLLM prior to 0.30.0

Identifiers

CVE-2026-105757
GHSA-85xf-c7hm-whqw
PYSEC-2026-4199

Timeline

2026-10-05: disclosed
2026-09-09: patched

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.