{"id":"CVE-2026-33625","summary":"LMDeploy vulnerable to arbitrary code execution via eval() of untrusted quant_dtype in model config loading","details":"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.","aliases":["GHSA-3hmm-rh5q-gwwr"],"modified":"2026-09-20T11:45:37.878167311Z","published":"2026-09-18T17:13:37.133Z","database_specific":{"osv_generated_from":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/33xxx/CVE-2026-33625.json","cna_assigner":"GitHub_M","cwe_ids":["CWE-400"]},"references":[{"type":"WEB","url":"https://github.com/InternLM/lmdeploy/releases/tag/v0.12.3"},{"type":"ADVISORY","url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/33xxx/CVE-2026-33625.json"},{"type":"ADVISORY","url":"https://github.com/InternLM/lmdeploy/security/advisories/GHSA-3hmm-rh5q-gwwr"},{"type":"ADVISORY","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-33625"}],"affected":[{"ranges":[{"type":"GIT","repo":"https://github.com/internlm/lmdeploy","events":[{"introduced":"e5df4e8336cd44a1c6e3e358ebd8c9b246932d99"},{"fixed":"8ea459f49ed9cd943481073011424919e31e3e3b"}],"database_specific":{"extracted_events":[{"introduced":"0.12.1"},{"fixed":"0.12.3"}],"source":"AFFECTED_FIELD"}}],"versions":["v0.12.2","v0.12.1"],"database_specific":{"source":"https://storage.googleapis.com/cve-osv-conversion/osv-output/CVE-2026-33625.json"}}],"schema_version":"1.9.0","severity":[{"type":"CVSS_V3","score":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H"}]}