Cyber Resilience

CVE-2025-62164

RCE in Vllm 0.10.2 – 0.11.1

Published
21 November 2025
Modified
04 December 2025
Patch / advisory
CVSS Score v3.1 8.8
Click a component to see what it means
Raw vectorCVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
EPSS Score 0.0092 57th percentile
Risk Priority 66 floored blend · peak EPSS

Summary

CVE-2025-62164 is a high-severity Improper Input Validation (CWE-20) vulnerability in Vllm Vllm. Its CVSS base score is 8.8 (High).

Operationally, exploitation aligns with the MITRE ATT&CK technique Exploitation for Privilege Escalation (T1068); ranked in the top 43% of CVEs by exploit likelihood; it is not currently listed in the CISA KEV catalog.

This vulnerability is AI-related — categorised as NLP and Transformers; in the Supply Chain and Deployment risk domain.

The strongest mitigations our analysis identified map to SA-11 (Developer Testing and Evaluation) and SI-10 (Information Input Validation) — see the control section below for these in your framework.

Deeper analysis AI-assisted summary

Synthesised by an AI model from the NVD description and linked references — a reading aid, not an authoritative source.

CVE-2025-62164 is a memory corruption vulnerability affecting vLLM, an inference and serving engine for large language models, in versions 0.10.2 through 0.11.0. The issue resides in the Completions API endpoint, which processes user-supplied prompt embeddings by loading serialized tensors via torch.load() without adequate validation. A change in PyTorch 2.8.0 disables sparse tensor integrity checks by default, allowing maliciously crafted tensors to bypass internal bounds checks and trigger an out-of-bounds memory write during the to_dense() call.

Attackers with low privileges (PR:L) can exploit this vulnerability over the network (AV:N) with low complexity (AC:L) and no user interaction (UI:N), as indicated by its CVSS v3.1 base score of 8.8. By submitting specially crafted prompt embeddings to the Completions API endpoint, an attacker can cause a denial-of-service crash or potentially achieve remote code execution on the hosting server, with high impacts on confidentiality, integrity, and availability (C:H/I:H/A:H). The vulnerability maps to CWEs including CWE-20 (Improper Input Validation), CWE-123 (Write-what-where Condition), CWE-502 (Deserialization of Untrusted Data), and CWE-787 (Out-of-bounds Write).

The vLLM project has patched this issue in version 0.11.1. Mitigation details are available in the project's security advisory (GHSA-mrw7-hf4f-83pf), the fixing pull request (#27204), and the commit (58fab50d82838d5014f4a14d991fdb9352c9c84b) that adds validation to prevent the exploitation of malformed sparse tensors.

This vulnerability is particularly relevant to AI/ML deployments, as vLLM is designed for serving LLMs, potentially exposing production inference servers to risks from untrusted inputs. No public reports of real-world exploitation are noted in the available information.

OWASP Top 10 for Web (2025)

EU & UK References

Vulnerability Data

vLLM is an inference and serving engine for large language models (LLMs). From versions 0.10.2 to before 0.11.1, a memory corruption vulnerability could lead to a crash (denial-of-service) and potentially remote code execution (RCE), exists in the Completions API endpoint.…

more

When processing user-supplied prompt embeddings, the endpoint loads serialized tensors using torch.load() without sufficient validation. Due to a change introduced in PyTorch 2.8.0, sparse tensor integrity checks are disabled by default. As a result, maliciously crafted tensors can bypass internal bounds checks and trigger an out-of-bounds memory write during the call to to_dense(). This memory corruption can crash vLLM and potentially lead to code execution on the server hosting vLLM. This issue has been patched in version 0.11.1.

CWE(s)

AI Security AnalysisAI

AI Category
NLP and Transformers
Risk Domain
Supply Chain and Deployment
OWASP Top 10 for LLMs 2025
None mapped
Classification Reason
Matched keywords: llms, pytorch, vllm

Related Threats

MITRE ATT&CK Enterprise Techniques

T1068 Exploitation for Privilege Escalation Privilege Escalation
Adversaries may exploit software vulnerabilities in an attempt to elevate privileges.
T1190 Exploit Public-Facing Application Initial Access
Adversaries may attempt to exploit a weakness in an Internet-facing host or system to initially access a network.
T1203 Exploitation for Client Execution Execution
Adversaries may exploit software vulnerabilities in client applications to execute code.
T1210 Exploitation of Remote Services Lateral Movement
Adversaries may exploit remote services to gain unauthorized access to internal systems once inside of a network.
T1211 Exploitation for Stealth Stealth
Adversaries may exploit vulnerabilities to evade detection by hiding activity, suppressing logging, or operating within trusted or unmonitored components.
T1212 Exploitation for Credential Access Credential Access
Adversaries may exploit software vulnerabilities in an attempt to collect credentials.
Derived from this CVE’s CWE(s) via the direct CWE→ATT&CK cross-walk.

CVEs Like This One

CVE-2024-42479Shared CWE-123, CWE-787
CVE-2024-23496Shared CWE-787
CVE-2024-21836Shared CWE-787
CVE-2024-21825Shared CWE-787
CVE-2024-23605Shared CWE-787
CVE-2024-21802Shared CWE-787
CVE-2024-21473Shared CWE-20, CWE-787
CVE-2024-53029Shared CWE-20, CWE-787
CVE-2023-20721Shared CWE-20, CWE-787
CVE-2024-53030Shared CWE-20, CWE-787

Affected Assets

vllm
vllm
0.11.1 · 0.10.2 — 0.11.1

Mitigating Controls

Control response

Prevent
Stop it (NIST 800-53)

Detect
Catch it (NIST detect / respond)

Harden
Shrink the surface (DISA STIG)
  • 10 hardening rules · 5 OS baselines
Validate
Prove the fix (OWASP ASVS)

Mitigating Controls (NIST 800-53 r5) AI

Developer testing and evaluation can discover missing input validation through analysis or test cases.

SI-10 directly requires validity checks on information inputs, structurally preventing improper or missing validation.

Memory-protection mechanisms block unauthorized writes to arbitrary locations even if a write-what-where primitive exists.

Requiring documented development standards and tools can embed input-validation practices into the engineering process.

Secure engineering principles require memory-safe coding and bounds checking that eliminate the root cause of write-what-where flaws.

Process isolation confines the blast radius of an arbitrary write so it cannot affect other domains.

Mitigating Controls (NIST CSF 2.0) AI

Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→CSF cross-walk (authority under review) — links open the control.

PR.PS-06 mostly match
prevents

Secure SDLC practices directly require and enforce input validation during development.

ID.RA-01 partial match
prevents

Vulnerability scanning and recording can discover out-of-bounds write flaws so they can be remediated.

PR.PS-02 partial match
prevents

Patching or replacing vulnerable software directly eliminates known instances of this coding weakness.

Mitigating Controls (ISO/IEC 27001:2022 Annex A) AI

Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→ISO cross-walk (authority under review) — links open the control.

finds

Testing against a defined set of requirements and using code review plus vulnerability scanning forces validation of inputs and handling of unanticipated conditions, reducing the chance that malformed data will be accepted.

prevents

Secure-coding guidelines and mandatory security testing (including code scans) compel developers to validate and sanitize inputs at design and implementation time, lowering the incidence of malformed or malicious data reaching downstream components.

prevents

Mandating input controls that include integrity checks and input validation ensures that untrusted data is examined before use, blocking the root cause of many injection and malformed-data weaknesses.

prevents

Security-by-design principles explicitly call for data validation and sanitization at every layer, reducing the chance that malformed or malicious input will be processed without scrutiny.

prevents

Requiring language-specific secure coding standards, peer review, SAST and documented mitigation of common programming errors forces validation of all inputs before they are trusted.

prevents

Change management processes help ensure security fixes for such weaknesses are properly deployed.

Hardening callouts derived

Configuration rules from DISA STIG baselines that bear on weaknesses of the type cited by this CVE. Each rule is shown with the relationship its mapping actually records, against the CWE it was authored against. Derived via CVE→CWE over `controls_xwalks` (authoritative rows only; rows rated `none` are excluded).

Oracle Linux 8 (1 rule)
  • V-248592 OL 8 must clear memory when it is freed to prevent use-after-free attacks. prevents CWE-123
RHEL 8 (2 rules)
  • V-230265 RHEL 8 must prevent the installation of software, patches, service packs, device drivers, or operating system components of local packages without verification they have been digitally signed using a certificate that is issued by a Certificate Authority (CA) that is recognized and approved by the organization. prevents CWE-20
  • V-230279 RHEL 8 must clear memory when it is freed to prevent use-after-free attacks. prevents CWE-123
Windows 10 (1 rule)
  • V-220727 Structured Exception Handling Overwrite Protection (SEHOP) must be enabled. prevents CWE-123
Windows 11 (1 rule)
  • V-253284 Structured Exception Handling Overwrite Protection (SEHOP) must be enabled. prevents CWE-123

References