CVE Feed

    Dashboard / CVE / CVE-2026-126397

    CVE-2026-126397

    vLLM from 0.10.1 before 0.31.0 forwards a request-controlled mm_processor_kwargs.max_patches value to the Hugging Face processor for Cohere2 Vision models without a server-side upper bound, causing quadratic CPU work during multimodal preprocessing. Remote clients can send small image chat-completion requests with large, varying max_patches values to bypass the LRU cache, exhaust preprocessing workers, and deny service to other clients.

    Published:Nov 13, 2026
    Last Modified:Nov 13, 2026
    EPS:Nov 13, 2026
    EPSS Score:0.00655
    CVSS Score:7.5

    Affected Products

    Vendor
    Vllm
    Product
    Vllm
    Vendor
    Vllm-project
    Product
    Vllm

    Exploits

    No exploit reference

    Related CVEs

    CVE-2026-126394 (CVSS 6.9) — vLLM 0.30.0 before 0.31.0 contains a resource exhaustion vulnerability in OpenCVDynamicOpenPanguVideoBackend, …
    CVE-2026-126395 (CVSS 6.9) — vLLM from 0.14.0rc1 before 0.31.0 contains an information disclosure vulnerability in sanitize_message() whose…
    CVE-2026-126396 (CVSS 6.9) — vLLM from 0.24.0 before 0.31.0 contains a resource exhaustion vulnerability in the MiniMax M3 VL video process…
    CVE-2026-126398 (CVSS 8.2) — vLLM from 0.28.0 before 0.30.0 contains a reachable assertion vulnerability in GPUModelRunner._init_xdrope_pos…
    CVE-2026-126399 (CVSS 7.1) — vLLM from 0.19.0 before 0.31.0 contains a resource exhaustion vulnerability that allows inference clients to b…
    CVE-2026-115918 (CVSS 6.5) — vLLM is an inference and serving engine for large language models. From 0.16.0 until 0.32.0, the /v1/responses…
    CVE-2026-115917 (CVSS 3.4) — vLLM is an inference and serving engine for large language models. Prior to 0.32.0, the OpenAI-compatible serv…
    CVE-2026-115916 (CVSS 6.5) — vLLM is an inference and serving engine for large language models. Prior to 0.32.0, vllm/tokenizers/hf.py comp…
    CVE-2026-115915 (CVSS 8.1) — vLLM is an inference and serving engine for large language models. Prior to 0.32.0, ImageMediaIO passes attack…
    CVE-2026-115914 (CVSS 6.5) — vLLM is an inference and serving engine for large language models. From 0.11.1 until 0.32.0, the DeepSeek-OCR,…

    Common Vulnerability Scoring System

    Attack Vector
    Network
    Adjacent
    Local
    Physical
    Privileges Required
    None
    Low
    High
    User Interaction
    None
    Required
    Scope
    Unchanged
    Changed
    Confidentiality
    None
    Low
    High
    Integrity
    None
    Low
    High
    Availability
    None
    Low
    High