CVE-2025-46570
published 2025-05-29CVE-2025-46570: vLLM is an inference and serving engine for large language models (LLMs). Prior to version 0.9.0, when a new prompt is processed, if the PageAttention…
PriorityP48low2.6CVSS 3.1
AVNACHPRLUIRSUCLINAN
EPSS
0.25%
16.1th percentile
vLLM is an inference and serving engine for large language models (LLMs). Prior to version 0.9.0, when a new prompt is processed, if the PageAttention mechanism finds a matching prefix chunk, the prefill process speeds up, which is reflected in the TTFT (Time to First Token). These timing differences caused by matching chunks are significant enough to be recognized and exploited. This issue has been patched in version 0.9.0.
Affected
4 ranges
| Vendor | Product | Version range | Fixed in |
|---|---|---|---|
| vllm-project | vllm | < 0.9.0 | 0.9.0 |
| vllm | vllm | < 0.9.0 | 0.9.0 |
| vllm | vllm | >= 0 < 0.9.0 | 0.9.0 |
| vllm | vllm | >= 0 < 77073c77bc2006eb80ea6d5128f076f5e6c6f54f | 77073c77bc2006eb80ea6d5128f076f5e6c6f54f |
CVSS provenance
nvdv3.12.6LOWCVSS:3.1/AV:N/AC:H/PR:L/UI:R/S:U/C:L/I:N/A:N
vendor_redhat2.6LOW
Stop checking back — get the weekly exploitation signal.
Every Monday: what got weaponized or added to CISA KEV in the last seven days — each CVE cross-linked to its PoC, Nuclei template, and detection rule. Free, one email a week, unsubscribe in one click.
OSV
CVE-2025-46570: vLLM is an inference and serving engine for large language models (LLMs)
osv·2025-05-29
CVE-2025-46570 CVE-2025-46570: vLLM is an inference and serving engine for large language models (LLMs)
vLLM is an inference and serving engine for large language models (LLMs). Prior to version 0.9.0, when a new prompt is processed, if the PageAttention mechanism finds a matching prefix chunk, the prefill process speeds up, which is reflected in the TTFT (Time to First Token). These timing differences caused by matching chunks are significant enough to be recognized and exploited. This issue has been patched in version 0.9.0.
GHSA
Potential Timing Side-Channel Vulnerability in vLLM’s Chunk-Based Prefix Caching
ghsa·2025-05-28
CVE-2025-46570 [LOW] CWE-208 Potential Timing Side-Channel Vulnerability in vLLM’s Chunk-Based Prefix Caching
Potential Timing Side-Channel Vulnerability in vLLM’s Chunk-Based Prefix Caching
This issue arises from the prefix caching mechanism, which may expose the system to a timing side-channel attack.
## Description
When a new prompt is processed, if the PageAttention mechanism finds a matching prefix chunk, the prefill process speeds up, which is reflected in the TTFT (Time to First Token). Our tests revealed that the timing differences caused by matching chunks are significant enough to be recognized and exploited.
For instance, if the victim has submitted a sensitive prompt or if a valuable system prompt has been cached, an attacker sharing the same backend could attempt to guess the victim's input. By measuring the TTFT based on prefix matches, the attacker could verify if their guess is
OSV
Potential Timing Side-Channel Vulnerability in vLLM’s Chunk-Based Prefix Caching
osv·2025-05-28
CVE-2025-46570 [LOW] Potential Timing Side-Channel Vulnerability in vLLM’s Chunk-Based Prefix Caching
Potential Timing Side-Channel Vulnerability in vLLM’s Chunk-Based Prefix Caching
This issue arises from the prefix caching mechanism, which may expose the system to a timing side-channel attack.
## Description
When a new prompt is processed, if the PageAttention mechanism finds a matching prefix chunk, the prefill process speeds up, which is reflected in the TTFT (Time to First Token). Our tests revealed that the timing differences caused by matching chunks are significant enough to be recognized and exploited.
For instance, if the victim has submitted a sensitive prompt or if a valuable system prompt has been cached, an attacker sharing the same backend could attempt to guess the victim's input. By measuring the TTFT based on prefix matches, the attacker could verify if their guess is
Red Hat
vllm: vLLM’s Chunk-Based Prefix Caching Vulnerable to Potential Timing Side-Channel
vendor_redhat·2025-05-29·CVSS 2.6
CVE-2025-46570 [LOW] CWE-208 vllm: vLLM’s Chunk-Based Prefix Caching Vulnerable to Potential Timing Side-Channel
vllm: vLLM’s Chunk-Based Prefix Caching Vulnerable to Potential Timing Side-Channel
vLLM is an inference and serving engine for large language models (LLMs). Prior to version 0.9.0, when a new prompt is processed, if the PageAttention mechanism finds a matching prefix chunk, the prefill process speeds up, which is reflected in the TTFT (Time to First Token). These timing differences caused by matching chunks are significant enough to be recognized and exploited. This issue has been patched in version 0.9.0.
A timing discrepancy flaw was found in vLLM, where a prefix match on a user prompt can reveal other user prompts. An attacker must have user-level access to the vLLM instance to exploit this vulnerabi
Mitigation: Mitigation for this issue is either not available or the currently avai
No detection rules found.
No public exploits indexed.
No writeups or analysis indexed.
2025-05-29
Published