Vulnerability Details CVE-2026-105758
vLLM is an inference and serving engine for large language models. From 0.24.0 until 0.30.0, the Qwen2VLVideoBackend and Qwen3VLVideoBackend classes accept request-level values for the media_io_kwargs.video.max_frames and media_io_kwargs.video.fps fields without enforcing server-side ceilings. An unauthenticated caller can submit these values to the /tokenize endpoint, causing the sampler to decode every frame selected from attacker-controlled video input, consume disproportionate frontend memory, and potentially terminate the API process before scheduling or admission control. The Rust frontend is not affected because it rejects the media_io_kwargs field. This issue is fixed in version 0.30.0.
Exploit prediction scoring system (EPSS) score
EPSS Score 0.003
EPSS Ranking 21.1%
CVSS Severity
CVSS v3 Score 5.3
References
Products affected by CVE-2026-105758
-
cpe:2.3:a:vllm:vllm:0.24.0
-
cpe:2.3:a:vllm:vllm:0.25.0
-
cpe:2.3:a:vllm:vllm:0.25.1
-
cpe:2.3:a:vllm:vllm:0.26.0
-
cpe:2.3:a:vllm:vllm:0.26.1
-
cpe:2.3:a:vllm:vllm:0.27.0
-
cpe:2.3:a:vllm:vllm:0.27.1
-
cpe:2.3:a:vllm:vllm:0.28.0
-
cpe:2.3:a:vllm:vllm:0.28.1
-
cpe:2.3:a:vllm:vllm:0.29.0
-
cpe:2.3:a:vllm:vllm:0.29.1