PT-2026-2260 · Vllm · Vllm

Oxcabe · Published 2026-01-10 · Updated 2026-01-13 · CVE-2026-22773

Positive Technologies (AI Summary)

CVSS v3.1

7.5

High

Base vector

AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Name of the Vulnerable Software and Affected Versions vLLM versions 0.6.4 through 0.11.9

Description vLLM is an inference and serving engine for large language models (LLMs). Users can cause the vLLM engine to crash when serving multimodal models that utilize the Idefics3 vision model implementation. This is achieved by submitting a specifically designed 1x1 pixel image. The crafted image triggers a tensor dimension mismatch, resulting in an unhandled runtime error and complete server termination.

Recommendations Update to version 0.12.0 or later.

Exploit

Fix

DoS

Allocation of Resources Without Limits

Weakness Enumeration

CWE-770

Related Identifiers

CVE-2026-22773

GHSA-GRG2-63FW-F2QR

PYSEC-2026-143

Affected Products

Vllm

References · 12

• 🔥 https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/22xxx/CVE-2026-22773.json · Exploit
• https://github.com/vllm-project/vllm/security/advisories/GHSA-grg2-63fw-f2qr · Vendor Advisory
• https://osv.dev/vulnerability/CVE-2026-22773 · Vendor Advisory
• https://osv.dev/vulnerability/PYSEC-2026-143 · Vendor Advisory
• https://osv.dev/vulnerability/GHSA-grg2-63fw-f2qr · Vendor Advisory
• https://nvd.nist.gov/vuln/detail/CVE-2026-22773 · Security Note
• https://github.com/vllm-project/vllm · Note
• https://github.com/vllm-project/vllm/commit/0ec84221718d920c3f46da879cc354f94b8fb59e · Note
• https://github.com/vllm-project/vllm/pull/29881 · Note
• https://t.me/cveNotify/148346 · Telegram Post
• https://t.me/CVEtracker/41932 · Telegram Post
• https://twitter.com/CVEnew/status/2009891366801740165 · Twitter Post