---
cve: "CVE-2026-73557"
severity: "MEDIUM"
cvss: 6.3
epss: "25%"
vendor: "vllm-project"
kev: false
exploited: false
published: "2026-08-13 15:20:18"
tags: [cve, security, medium]
source: tsecurity.de CVE-Dossier
exported: "2026-09-19T19:02:28+02:00"
---

# CVE-2026-73557

> 6.3 MEDIUM · 🧪 PoC

## Beschreibung

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.

## CVSS-Vektor

```
CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N
```

| Metrik | Wert | Bewertung |
|---|---|---|
| AV Angriffsvektor | Netzwerk | bad |
| AC Komplexität | Gering | bad |
| PR Privilegien | Keine | bad |
| UI Interaktion | Keine | bad |

## Patch verfügbar (OSV)

- 793cf79c89d4049124e756915468ac30318f2e50 (Commit)
- ffd46bfab2128bb84146050e98b51a617c6575ab (Commit)

## Referenzen

- <https://github.com/vllm-project/vllm/security/advisories/GHSA-pr7f-p5mw-fc87>
- <https://github.com/vllm-project/vllm/pull/48583>
- <https://github.com/vllm-project/vllm/commit/793cf79c89d4049124e756915468ac30318f2e50>
- <https://github.com/vllm-project/vllm/releases/tag/v0.26.0>

---
_Exportiert aus dem [tsecurity.de CVE-Dossier](https://tsecurity.de/cve?cve=CVE-2026-73557) · Datenquellen: EUVD (ENISA), NVD, OSV, CISA KEV, FIRST EPSS, Exploit-DB, BSI BITS_