Threats Tagged 'cve-2026-34755'

Red Hat® Enterprise Linux® AI is a foundation model platform to seamlessly develop, test, and run Granite family large language models (LLMs) for enterprise applications. This update provides the latest Red Hat Enterprise Linux AI 3.4.4 container disk images for use with OpenShift Virtualization. For a full list of changes in this release, see the Red Hat Enterprise Linux AI Release Notes linked in the References section.

Red Hat AI Inference

Red Hat AI Inference

Red Hat® AI Inference Server Model Optimization Tools

CVE-2026-26740 is a high-severity vulnerability affecting Red Hat AI Inference Server 3.2.2 (CUDA). It involves a security issue categorized under multiple CWEs including out-of-bounds write (CWE-787) and others. The vulnerability impacts the integrity and availability of the affected product. No official patch or fix is currently provided by Red Hat. There are no known exploits in the wild at this time. The vendor advisory confirms the availability of the updated Red Hat AI Inference Server 3.2.2 (CUDA) but does not specify a fix for this CVE. The affected versions include Red Hat AI Inference Server 3.2 and related container images. No geographic targeting is indicated.

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.