CVE-2026-53923
Fixed in a distro release EPSS <1%CVSS 5.3
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.
Products: vllm-project vllm
Fixed upstream
| Ecosystem | Package | Fixed in |
|---|---|---|
| PyPI | vllm | 0.24.0 |
| PyPI | vllm | no fix |
Distro releases
| Release | Package | Fixed in |
|---|---|---|
| Chainguard | lmcache-cuda-12.8 | no fix |
| Chainguard | nemo-rl-cuda-13.0-vllm | no fix |
| Chainguard | py3.10-vllm-cuda-12.4 | no fix |
| Chainguard | py3.10-vllm-cuda-12.9 | no fix |
| Chainguard | py3.10-vllm-cuda-13.0 | no fix |
| Chainguard | py3.12-vllm-cuda-12.4 | no fix |
| Chainguard | py3.12-vllm-cuda-12.9 | no fix |
| Chainguard | py3.12-vllm-cuda-13.0 | no fix |
| Chainguard | py3.13-vllm-cuda-12.9 | no fix |
| Chainguard | py3.13-vllm-cuda-13.0 | no fix |
| Chainguard | tritonserver-backend-vllm-cuda-12.9 | no fix |
| Chainguard | tritonserver-backend-vllm-cuda-13.0 | 25.11-r12 |
| Chainguard | vllm-cuda-13.2 | 0.24.0-r0 |
| Chainguard | vllm-openai-cuda-12.9 | 0.28.0-r0 |
| Chainguard | vllm-openai-cuda-12.9-nvidia-cusparse | no fix |
| Chainguard | vllm-openai-cuda-13.0 | 0.24.0-r1 |