CVE-2026-34755

vLLM is an inference and serving engine for large language models (LLMs). From 0.7.0 to before 0.19.0, the VideoMediaIO.load_base64() method at vllm/multimodal/media/video.py splits video/jpeg data URLs by comma to extract individual JPEG frames, but does not enforce a frame count limit. The num_frames parameter (default: 32), which is enforced by the load_bytes() code path, is completely bypassed in the video/jpeg base64 path. An attacker can send a single API request containing thousands of comma-separated base64-encoded JPEG frames, causing the server to decode all frames into memory and crash with OOM. This vulnerability is fixed in 0.19.0.

The following products are affected by CVE-2026-34755 vulnerability.

Here, you will find a curated list of external links that provide in-depth information to CVE-2026-34755.

URL	Resource
https://github.com/vllm-project/vllm/security/advisories/GHSA-pq5c-rjhq-qp7p
https://nvd.nist.gov/vuln/detail/CVE-2026-34755
https://www.cve.org/CVERecord?id=CVE-2026-34755