Vulnerability · Published Apr 20, 2026 · Assigned by GitHub_M

CVE-2026-33626LMDeploy Vulnerable to Server-Side Request Forgery (SSRF) via Vision-Language Image Loading

High 7.5

In plain English

A high-severity server-side request forgery in lmdeploy from InternLM. Proof-of-concept exploit code is public. The chance it is exploited in the next 30 days is estimated at 45.2% (higher than 99% of all known vulnerabilities). Ready-made attack tooling exists: Nuclei, public proof-of-concept code.

Severity (CVSS 3.1)7.5High

Exploit odds (30 days)45.2%top 1.2% of all CVEs

Exploited in the wildNo reportsCISA SSVC: poc

Attack tooling2Nuclei, Public PoC

Official description

LMDeploy is a toolkit for compressing, deploying, and serving large language models. Versions prior to 0.12.3 have a Server-Side Request Forgery (SSRF) vulnerability in LMDeploy's vision-language module. The `load_image()` function in `lmdeploy/vl/utils.py` fetches arbitrary URLs without validating internal/private IP addresses, allowing attackers to access cloud metadata services, internal networks, and sensitive resources. Version 0.12.3 patches the issue.

Affected packages & fixed versions

Ecosystem	Package	Fixed in	Advisory
PyPI	lmdeploy	no fix listed	GHSA-6w67-hwm5-92mq
PyPI	lmdeploy	no fix listed	PYSEC-2026-2607

Affected products (as reported)

Vendor	Product	Versions
InternLM	lmdeploy	< 0.12.3

Weakness: CWE-918 · Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N

References

https://github.com/InternLM/lmdeploy/security/advisories/GHSA-6w67-hwm5-92mq x_refsource_CONFIRM
https://github.com/InternLM/lmdeploy/pull/4447 x_refsource_MISC
https://github.com/InternLM/lmdeploy/commit/71d64a339edb901e9005358e0633fbbab367d626 x_refsource_MISC
https://github.com/InternLM/lmdeploy/releases/tag/v0.12.3 x_refsource_MISC