CVE-2026-22773: vLLM engine crash via crafted 1x1 pixel image in Idefics3 multimodal models

Summary

CVE-2026-22773 affects vLLM, an inference and serving engine for large language models, in versions from 0.6.4 before 0.12.0. A specially crafted 1x1 pixel image sent to a server running multimodal models that use the Idefics3 vision model implementation causes a tensor dimension mismatch. The resulting unhandled runtime error terminates the whole server.

Mitigation

This issue has been patched in version 0.12.0.

Affected packages in the Exposure Registry

Matched by package name and ecosystem. Each entry shows whether the package delegates to a language model and how many tracked packages depend on it.

• vllmPyPILLM dependency since 2023-08-25 · 15 tracked dependents

Topics

• Inference infrastructure

Related items

• LowCloakFEM: Agentic FEM Code Generation for Red-Teaming Computational Resource ExhaustionSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
• MediumThe Rise of LLMJacking: How Threat Actors Monetize Compromised Cloud and Model ResourcesSimilar attack · Google Threat Intelligence
• InfoThe Blind Spots of Detection: Adversarial Perturbation Attacks against Watchdogs in Distributed LLM InferenceSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
• HighCVE-2026-118761: Kubeflow Pipelines decompression bomb in pipeline archive uploadSimilar attack · NVD/CVE Database
• LowOpenCode User Hit with Hostile Backlash After Reporting Free AI LoopholeSimilar attack · Socket Blog