Vulnerability Description
vLLM is an inference and serving engine for large language models (LLMs). From 0.5.5 until 0.23.1rc0, integer truncation of tensor dimensions in vLLM's GGUF dequantize kernels (csrc/quantization/gguf/gguf_kernel.cu) causes partial tensor processing. The output tensor is allocated at full size via torch::empty (uninitialized memory), but the dequantize CUDA kernel processes only a truncated number of elements. The unfilled portion of the output tensor retains whatever was previously in GPU memory. In multi-tenant inference deployments, this residual GPU memory may contain tensor data from other users' inference requests, constituting information disclosure. This vulnerability is fixed in 0.23.1rc0.
CVSS Score
HIGH
Affected Products
| Vendor | Product | Versions |
|---|---|---|
| Vllm | Vllm | >= 0.5.5, < 0.23.1 |
Related Weaknesses (CWE)
References
- https://github.com/vllm-project/vllm/commit/f219788f91952827132fa4fdf916427cd20dPatch
- https://github.com/vllm-project/vllm/pull/44971Issue Tracking
- https://github.com/vllm-project/vllm/security/advisories/GHSA-5jv2-g5wq-cmr4Third Party Advisory
FAQ
What is CVE-2026-53923?
CVE-2026-53923 is a vulnerability with a CVSS score of 7.5 (HIGH). vLLM is an inference and serving engine for large language models (LLMs). From 0.5.5 until 0.23.1rc0, integer truncation of tensor dimensions in vLLM's GGUF dequantize kernels (csrc/quantization/gguf/...
How severe is CVE-2026-53923?
CVE-2026-53923 has been rated HIGH with a CVSS base score of 7.5/10. Review the CVSS metrics above for detailed severity breakdown.
Is there a patch for CVE-2026-53923?
Check the references section above for vendor advisories and patch information. Affected products include: Vllm Vllm.