vLLM Denial of Service via the best_of parameter
Medium6.2CVE-2024-8939 · Published Sep 17, 2024 · updated Aug 7, 2026
A vulnerability was found in the ilab model serve component, where improper handling of the best_of parameter in the vllm JSON web API can lead to a Denial of Service (DoS). The API used for LLM-based sentence or chat completion accepts a best_of parameter to return the best completion from several options. When this parameter is set to a large value, the API does not handle timeouts or resource exhaustion properly, allowing an attacker to cause a DoS by consuming excessive system resources. This leads to the API becoming unresponsive, preventing legitimate users from accessing the service.
Affected versions
| Package | Affected | Fixed in |
|---|---|---|
| vllm PyPI | <= 0.5.0.post1 | No fix yet |
Details and references
- CVSS 3.1
- CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
- Severity from
- GitHub (reviewed advisory)
- Weakness
- CWE-400
- Also known as
- CVE-2024-8939, PYSEC-2026-2025
More vLLM advisories
All vLLM| Date | Advisory | Severity | Fixed in |
|---|---|---|---|
| Mar 202025 | vLLM Deserialization of Untrusted Data vulnerability | Critical9.8 | No fix yet |
| Mar 192025 | vLLM Allows Remote Code Execution via Mooncake Integration | Critical9.0 | 0.8.0 |
| Mar 192025 | vLLM denial of service via outlines unbounded cache on disk | Medium6.5 | 0.8.0 |
| Feb 62025 | vLLM uses Python 3.12 built-in hash() which leads to predictable hash collisions in prefix cache | Low2.6 | 0.7.2 |
| Jan 272025 | vllm: Malicious model to RCE by torch.load in hf_model_weights_iterator | High7.5 | 0.7.0 |
| Sep 172024 | vLLM denial of service vulnerability | High7.5 | 0.5.5 |