vllm: Inconsistent Interpretation of HTTP Requests
CVE-2026-48746 is a critical-severity inconsistent interpretation of HTTP requests vulnerability in vllm from Vllm-project, with a CVSS 3.1 base score of 9.1 out of 10. It can be exploited remotely over the network by an unauthenticated attacker, without any action from a user. Affected vllm versions: >= 0.3.0, < 0.22.0. The record does not name a fixed version, so check the vendor advisory. There is no confirmed exploitation so far. EPSS estimates a 1.15% chance of exploitation within 30 days, higher than 67% of scored CVEs.
Critical severity: a successful attack would have severe impact.
What to do
- Check which version of vllm you run against the affected versions below, and install the vendor's latest release.
- Read the advisory on github.com for workarounds and exact fixed builds.
- Until you can update, restrict network access to the affected service so it is not reachable from untrusted networks.
Affected products and versions
| Vendor | Product | Affected versions | Fixed in |
|---|---|---|---|
| Vllm-project | vllm | >= 0.3.0, < 0.22.0 | Not stated |
| Red Hat | Red Hat AI Inference Server 3.2 | Not specified in the record | Not stated |
| Red Hat | Red Hat AI Inference Server 3.3 | Not specified in the record | Not stated |
| Red Hat | Red Hat Ansible Automation Platform 2.6 | Not specified in the record | Not stated |
| Red Hat | Red Hat Ansible Automation Platform 2.7 | Not specified in the record | Not stated |
| Red Hat | Red Hat Migration Toolkit for Applications 8.2 | Not specified in the record | Not stated |
| Red Hat | Red Hat OpenShift AI 2.25 | Not specified in the record | Not stated |
| Red Hat | Exploit Intelligence | Not specified in the record | Not stated |
| Red Hat | OpenShift Lightspeed | Not specified in the record | Not stated |
| Red Hat | Red Hat AI Inference Server | Not specified in the record | Not stated |
| Red Hat | Red Hat Ansible Automation Platform 2 | Not specified in the record | Not stated |
| Red Hat | Red Hat Enterprise Linux AI (RHEL AI) 3 | Not specified in the record | Not stated |
| Red Hat | Red Hat Hardened Images | Not specified in the record | Not stated |
| Red Hat | Red Hat OpenShift AI (RHOAI) | Not specified in the record | Not stated |
| Red Hat | Red Hat Satellite 6 | Not specified in the record | Not stated |
Version data comes from the organisation that assigned the CVE. Confirm against the vendor advisory before deciding you are unaffected.
How the severity score breaks down
- Attack vector
- NetworkCan be exploited remotely over a network, including the internet.
- Attack complexity
- LowNo special conditions are needed; the attack works reliably.
- Privileges required
- NoneNo account or login is needed.
- User interaction
- NoneNo action by a victim is required.
- Scope
- UnchangedImpact stays within the vulnerable component.
- Confidentiality
- HighAll data handled by the component can be read.
- Integrity
- NoneNothing can be modified.
- Availability
- HighThe component can be taken down completely.
Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H. Scored by CNA (github.com).
| Scoring system | Score | Severity | Scored by |
|---|---|---|---|
| CVSS 3.1 | 9.1 | Critical | CNA (github.com) |
| CVSS 3.1 | 9.1 | Critical | CNA |
Official description
vLLM is an inference and serving engine for large language models (LLMs). From 0.3.0 until 0.22.0, a vulnerability in ASGI web servers and starlette's trust on those web servers enables an authentication bypass of the OpenAI API AuthenticationMiddleware. It allows to use the API without providing the configured VLLM_API_KEY or --api-key. This vulnerability is fixed in 0.22.0.
Source: the CVE record as published by its numbering authority, via NVD and CVE.org.
The underlying weakness: Inconsistent Interpretation of HTTP Requests
The product acts as an intermediary HTTP agent (such as a proxy or firewall) in the data flow between two entities such as a client and server, but it does not interpret malformed HTTP requests or responses in ways that are consistent with how the messages will be processed by those entities that are at the ultimate destination.
How developers prevent this class of flaw
- Use a web server that employs a strict HTTP parsing procedure, such as Apache.
- Terminate the client session after each request.
More Inconsistent Interpretation of HTTP Requests vulnerabilities (CWE-444)
Timeline
- CVE record published.
- CISA published its SSVC assessment (exploitation: poc, automatable: yes, technical impact: partial).
- Record last changed materially (score, affected products, status or references).
References
Advisories
- github.com/vllm-project/vllm/security/advisories/GHSA-94f4-hr76-p5j6Third Party Advisory
Patches and release notes
- github.com/vllm-project/vllm/pull/43426Issue Tracking
Third-party and government advisories
- x41-dsec.de/lab/advisories/x41-2026-002-starletteThird Party Advisory
Other references
- access.redhat.com/errata/RHSA-2026:30088
- access.redhat.com/errata/RHSA-2026:30089
- access.redhat.com/errata/RHSA-2026:36005
- access.redhat.com/errata/RHSA-2026:36006
- access.redhat.com/errata/RHSA-2026:42132
- access.redhat.com/errata/RHSA-2026:42142
- access.redhat.com/errata/RHSA-2026:42644
- access.redhat.com/errata/RHSA-2026:43038
- access.redhat.com/errata/RHSA-2026:61627
- access.redhat.com/errata/RHSA-2026:61629
- access.redhat.com/security/cve/CVE-2026-48746
- bugzilla.redhat.com/show_bug.cgi?id=2491581
- security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-48746.json
Other vulnerabilities in vllm
- 7.5 High severity
- 9.8 Critical severity
- 8.8 High severity
- 7.7 High severity
- 7.5 High severity
- 7.5 High severity
Recent Inconsistent Interpretation of HTTP Requests vulnerabilities
- 9.1 Critical severity
- 7.5 High severity
- 9.1 Critical severity
- 9.1 Critical severity
- 6.5 Medium severity
- 7.5 High severity
Spotted an error? Tell us.