PT-2025-53604 · Lmdeploy · Lmdeploy

Yueyuel · Published 2025-12-26 · Updated 2025-12-27 · CVE-2025-67729

Positive Technologies (AI Summary)

CVSS v3.1

8.8

High

Base vector

AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

Name of the Vulnerable Software and Affected Versions LMDeploy versions prior to 0.11.1

Description LMDeploy is a toolkit used for compressing, deploying, and serving LLMs. A flaw exists where the torch.load() function is called without the weights only=True parameter when loading model checkpoint files. This insecure deserialization allows an attacker to execute arbitrary code on a system by loading a malicious .bin or .pt model file. The torch.load() function is used to load serialized objects, and without the weights only=True parameter, it can deserialize arbitrary Python objects, leading to code execution.

Recommendations LMDeploy versions prior to 0.11.1 are vulnerable and should be updated to version 0.11.1 or later.

Exploit

Fix

RCE

Deserialization of Untrusted Data

Weakness Enumeration

CWE-502

Related Identifiers

CVE-2025-67729

GHSA-9PF3-7RRR-X5JH

Affected Products

Lmdeploy

References · 11

• 🔥 https://github.com/InternLM/lmdeploy/security/advisories/GHSA-9pf3-7rrr-x5jh · Exploit
• 🔥 https://github.com/CVEProject/cvelistV5/tree/main/cves/2025/67xxx/CVE-2025-67729.json · Exploit
• https://github.com/InternLM/lmdeploy/commit/eb04b4281c5784a5cff5ea639c8f96b33b3ae5ee · Patch
• https://nvd.nist.gov/vuln/detail/CVE-2025-67729 · Security Note
• https://osv.dev/vulnerability/CVE-2025-67729 · Vendor Advisory
• https://t.me/cveNotify/146920 · Telegram Post
• https://t.me/CVEtracker/40729 · Telegram Post
• https://twitter.com/CVEnew/status/2004868112437707037 · Twitter Post
• https://twitter.com/TheHackerWire/status/2004678347411128754 · Twitter Post
• https://twitter.com/VulmonFeeds/status/2004684009310244948 · Twitter Post
• https://twitter.com/CveFindCom/status/2004683411563147563 · Twitter Post