Skip to content

CVE-2026-33625 LMDeploy vulnerable to arbitrary code execution via eval() of untrusted quant_dtype in model config loading

HighCVSS 8.8 EPSS 0.44%
Published
Updated
Exploit status
No public exploit known
CISA SSVC
Proof of concept public · automatable: no · technical impact: total

Description

LMDeploy is a toolkit for compressing, deploying, and serving large language models. Versions 012.1 through 0.12.2 contain a code injection vulnerability in lmdeploy/pytorch/config.py line 620 that allows an attacker to execute arbitrary Python code by publishing a malicious HuggingFace model with a crafted quantization_config.quant_dtype value. When a user loads the model with lmdeploy, the quant_dtype is passed to eval(f'torch.{quant_dtype}') without any validation. Version 0.12.3 contains a patch.

Upgrade path

2 packages: 1 fixed.

PackageAffectedFixed inStatus
pypi lmdeploy(PyPI) = 0.12.1; = 0.12.2; >= 0.12.1, < 0.12.3 0.12.3 Fixed
InternLM lmdeploy = >= 0.12.1, < 0.12.3 No fix listed Versions listed

Affected products

ProductAffected versionsFixed in
pypi lmdeploy(PyPI) >= 0.12.1, < 0.12.3 0.12.3
pypi lmdeploy(PyPI) = 0.12.1 No fix listed
pypi lmdeploy(PyPI) = 0.12.2 No fix listed
InternLM lmdeploy = >= 0.12.1, < 0.12.3 No fix listed

References