Vulnerability in Open Neural Network Exchange Affecting Multiple Versions
CVE-2026-115921

6.2MEDIUM

Key Information:

Vendor

Onnx

Status
Onnx
Vendor
CVE Published:
26 October 2026

What is CVE-2026-115921?

Open Neural Network Exchange (ONNX) versions 1.4.0 through 1.22.0 contain a vulnerability where the attributes required for the ScanInferenceFunction are not properly validated. This flaw can lead to dereferencing of a null pointer if certain attributes are missing, or oversized memory allocations when negative values are mishandled. Consequently, untrusted models can inadvertently disrupt shape inference processes, raising significant security concerns. This issue was resolved in version 1.22.0, encouraging users to update promptly.

Affected Version(s)

onnx >= 1.4.0, < 1.22.0

References

https://github.com/onnx/onnx/security/advisories/GHSA-m4p...
x_refsource_CONFIRM
https://github.com/onnx/onnx/pull/7916
x_refsource_MISC
https://github.com/onnx/onnx/commit/dd6bc92ea39fd2274fd67...
x_refsource_MISC

CVSS V3.1

Score:
6.2
Severity:
MEDIUM
Confidentiality:
None
Integrity:
None
Availability:
None
Attack Vector:
Local
Attack Complexity:
Low
Privileges Required:
None
User Interaction:
None
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.