CVE-2025-24357 — vLLM allows a malicious model RCE by torch.load in hf_model_weights_iterator

2026-10-15T03:22:25.803866+00:00

Affected: vllm-project vllm

vLLM is a library for LLM inference and serving. vllm/model_executor/weight_utils.py implements hf_model_weights_iterator to load the model checkpoint, which is downloaded from huggingface. It uses the torch.load function and the weights_only parameter defaults to False. When torch.load loads malicious pickle data, it will execute arbitrary code during unpickling. This vulnerability is fixed in v0.7.0.

GHSA-rh4j-5rhw-hr54 — vllm: Malicious model to RCE by torch.load in hf_model_weights_iterator

2026-10-15T03:22:25.803928+00:00

Affected: PyPI: vllm

### Description The vllm/model_executor/weight_utils.py implements hf_model_weights_iterator to load the model checkpoint, which is downloaded from huggingface. It use torch.load function and weights_only parameter is default value False. There is a security warning on https://pytorch.org/docs/stable/generated/torch.load.html, when torch.load load a malicious pickle data it will execute arbitrary code during unpickling.

### Impact This vulnerability can be exploited to execute arbitrary codes and OS commands in the victim machine who fetch the pretrained repo remotely.

Note that most models now use the safetensors format, which is not vulnerable to this issue.

### References * https://pytorch.org/docs/stable/generated/torch.load.html * Fix: https://github.com/vllm-project/vllm/pull/12366

Most recent entries from all

CVE-2025-24357 — vLLM allows a malicious model RCE by torch.load in hf_model_weights_iterator

GHSA-rh4j-5rhw-hr54 — vllm: Malicious model to RCE by torch.load in hf_model_weights_iterator