CVE-2025-23335
Official description Straight from the sourceThe vendor's or NVD's own wording, published unedited. Authoritative, but often terse — it says what broke, rarely what to do.
NVD · uneditedNVIDIA Triton Inference Server for Windows and Linux and the Tensor RT backend contain a vulnerability where an attacker could cause an underflow by a specific model configuration and a specific input. A successful exploit of this vulnerability might lead to denial of service.
Technical summary Written by usOur analysis, written from the advisory, the CVSS vector and the affected-version data. It adds context the advisory leaves out, and never invents facts that are not in the source.
dbcve analysis · moderate confidenceNVIDIA Triton Inference Server and TensorRT backend contain an underflow vulnerability triggered by specific model configurations and inputs. The underflow condition during inference can cause the service to fail, resulting in denial of service. This is a numerical edge case in floating-point operations within the ML inference pipeline.
Verify against the referenced sources before acting — the references below are authoritative for this CVE, this summary is not.
Affected products & versions What the vendor confirmedThe version ranges the vendor confirmed as vulnerable. If your version sits inside a range here, treat yourself as exposed until you have upgraded.
NVD · CPE data< 25.05CVSS breakdown How the score is builtThe industry scoring standard. It rates how the flaw is reached, what it takes to exploit, and what an attacker gains — the score is derived from those, not the other way round.
From the vector- Attack vector
- Network
- Complexity
- Low
- Privileges
- None
- User interaction
- None
- Scope
- Unchanged
- Confidentiality
- None
- Integrity
- None
- Availability
- High
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Am I affected? How to checkSteps we derive from the advisory and the affected-version data, so you can decide whether this CVE reaches your setup. They are a guide, not a scan — your own configuration is the authority.
dbcve checksWork through these to decide whether this CVE applies to you.
-
Identify Triton Inference Server versionRun 'tritonserver --version' or check the container image tag. If using a container orchestrator, inspect the deployed image version.Affected if Version is earlier than 25.05 (e.g., 25.03, 24.xx, etc.)
-
Confirm TensorRT backend is in useReview the Triton model repository configuration and check if any models are configured to use the TensorRT backend (model config 'platform' or 'backend' set to 'tensorrt').Affected if TensorRT backend is enabled and serving models in the inference pipeline;
-
Identify deployed model configurationsExamine model configuration files (config.pbtxt) in the model repository, particularly looking for floating-point parameter settings, quantization configs, or precision settings that may involve edge-case numerical operations.Affected if Model configs contain floating-point precision settings that could trigger underflow conditions;
-
Review inference input patternsInspect inference request logs or monitoring data for inputs that produce very small floating-point values, denormal numbers, or values near the minimum representable float.Affected if Inference requests include inputs that historically trigger numerical underflow in ML inference pipelines.
You are affected if running Triton Inference Server version below 25.05 with TensorRT backend and serving models that process inputs capable of triggering floating-point underflow during inference.
Generated from the published advisory. Verify against your own configuration.
Remediation Closing itWhat it takes to close this. Where a vendor fix exists we point at it; where none exists we say so plainly, and can build one. Effort estimates are scoped from the advisory, not from your codebase.
dbcve · scoped25.05
Apply NVIDIA's patches when available; review and validate model configurations used in Triton Server deployments; implement input validation to reject inputs that trigger the underflow condition.
Triton Inference Server 25.05
- 1. Back up your current Triton Inference Server configuration and models
- 2. Download Triton Inference Server version 25.05 or later from NVIDIA's official repository or container registry
- 3. Stop the currently running Triton Inference Server instance
- 4. Install the updated Triton Inference Server 25.05 package or pull the updated container image
- 5. Verify the model configuration and inputs that triggered the underflow are no longer vulnerable
- 6. Start the upgraded Triton Inference Server service
- 7. Test your inference workflows to confirm normal operation
Generated from the published advisory — verify against the referenced sources before acting.
- Consultation6.0 h
- Implementation12.0 h
- Testing10.0 h
- Review / QA6.0 h
An estimate, not a bill — we confirm scope with you before any work starts. Need it this week? Rush from $9,504.
Scan for this in your stack
Free · runs locallyCheck whether your project pulls in CVE-2025-23335 — or any other known-vulnerable package — straight from your lock files. Free and open source; it runs locally and uploads nothing.
References Go to the primary sourcePrimary sources — vendor advisories, patches and trackers. Where our summary and a reference disagree, the reference wins.
Primary sourcesPractitioner notes
ContributedPeer-ranked notes from engineers who’ve handled CVE-2025-23335 in production — separate from our analysis above.
The advisory tells you what broke. It rarely tells you what actually worked. If you’ve dealt with this one, that detail is what the next engineer is searching for.
- The version that genuinely resolved it — not the one the vendor claimed
- A config change or rule that shut the vector down
- A gotcha in the upgrade path that cost you an afternoon
No notes yet
Be the first to add a field note for this CVE — a mitigation you’ve verified, a version caveat, or a link to a working fix. Sign in above to contribute.
A place for practitioners to share what actually worked: a mitigation you’ve tested, a configuration change, a version- or environment-specific caveat, or a link to a verified patch. The most useful notes rise to the top as peers upvote them, so the signal stays high.
- Verified mitigations, workarounds, and config changes
- Version or environment caveats, and links to real fixes
- No weaponised exploit code, or anything meant to cause harm
- No spam, self-promotion, credentials, or personal data