Skip to content
LLMHigh

CVE-2026-73557: vLLM is an inference and serving engine for large language models. From 0.20.2rc0 until 0.26.0, safe…

Occurred
Aug 13, 2026
Published
Aug 13, 2026
Source
NVD CVE

Summary

vLLM is an inference and serving engine for large language models. From 0.20.2rc0 until 0.26.0, safe_load_prompt_embeds in vllm/renderers/embed_utils.py uses torch.sparse.check_sparse_tensor_invariants, whose process-global save, enable, and restore state can be raced by concurrent prompt_embeds parts submitted to POST /v1/chat/completions through AsyncMultiModalItemTracker.resolve_items, asyncio.gather, and the default executor, allowing an invalid sparse tensor to reach tensor.to_dense despite the CVE-2025-62164 guard when enable_prompt_embeds is enabled. This issue is fixed in version 0.26.0.

Tags

  • CVE
  • CVE-2026-73557
  • cvss:MEDIUM

Classification

Category confidence: 95%
Severity confidence: 90%

Rationale: CVE describes a race condition vulnerability in vLLM inference engine allowing invalid tensor state, representing a significant security flaw in LLM serving infrastructure.

Classifier: Qwen/Qwen3.5-35B-A3B-FP8

Visit source