vLLM: VLLM versions >= 0.10.2 and < 0.13.0 are missing sparse tensor validation in multimodal…
vLLM
More for vLLM RSS feed for vLLM
- Severity
- High, CVSS 7.5
- CVSS vector
- CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
- Record
- CVE-2026-56340
- Affects
- vLLM
- Known exploited
- Not listed in the CISA Known Exploited Vulnerabilities catalog
- Published
- Record first published by the source
- June 20, 2026
- Source last modified
- October 23, 2026
Description
vLLM versions >= 0.10.2 and < 0.13.0 are missing sparse tensor validation in multimodal embeddings processing. Because PyTorch disables sparse tensor invariant checks by default, an attacker can submit crafted embedding requests with malformed (negative or out-of-bounds) tensor indices, when the prompt-embeds feature is enabled, to trigger crashes or resource exhaustion (denial of service), with potential for out-of-bounds/write-what-where memory corruption. This continues CVE-2025-62164, whose prior fix only disabled the feature by default rather than addressing the root cause.
Description quoted from the NVD record.
Who is affected
vLLM is software that runs AI language models on a server so that other apps and people can send it questions and get answers. This flaw affects vLLM versions 0.10.2 up to, but not including, 0.13.0, and only when the optional feature that accepts prepared prompt data (called prompt embeds) is turned on. In those setups, an attacker could send specially made requests that crash the software or use up its resources, making it unavailable, and could possibly corrupt its memory. This flaw is rated High.
What to do
- The person who installed or runs vLLM for your organization should check which version is in use. If it is from
0.10.2up to, but not including,0.13.0, your setup is in the affected range. - Ask that person whether the prompt embeds feature is turned on, since this flaw only affects setups where it is on.
- No fixed version is named for this flaw yet, so check the vendor advisory linked on this page for a fixed release and install it.
- If an IT provider or vendor runs vLLM for you, send them this advisory and ask when the update will be done.
Analysis written from the source record as of October 23, 2026 and reviewed before publishing.