CVE-2026-24151
Published: 24 March 2026
Summary
CVE-2026-24151 is a high-severity Deserialization of Untrusted Data (CWE-502) vulnerability in Nvidia Megatron-Lm. Its CVSS base score is 7.8 (High).
Operationally, exploitation aligns with the MITRE ATT&CK technique Exploitation for Client Execution (T1203); ranked at the 22.3th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog.
The strongest mitigations our analysis identified are NIST 800-53 SI-10 (Information Input Validation) and SI-2 (Flaw Remediation).
Deeper analysis
CVE-2026-24151 is a vulnerability in NVIDIA Megatron-LM, specifically within its inferencing component. An attacker can cause remote code execution (RCE) by convincing a user to load a maliciously crafted input. The issue stems from CWE-502 (Deserialization of Untrusted Data) and carries a CVSS v3.1 base score of 7.8 (AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H), indicating high impact with local access, low complexity, and low privileges required.
A local attacker with low privileges can exploit this vulnerability by socially engineering a user to process the malicious input during Megatron-LM inferencing. Successful exploitation enables arbitrary code execution, privilege escalation, information disclosure, and data tampering on the affected system.
Mitigation details are available in official advisories, including the NVIDIA security bulletin at https://nvidia.custhelp.com/app/answers/detail/a_id/5769, the NVD entry at https://nvd.nist.gov/vuln/detail/CVE-2026-24151, and the CVE record at https://www.cve.org/CVERecord?id=CVE-2026-24151.
This vulnerability affects an AI/ML framework used for large language model inferencing, underscoring risks associated with untrusted inputs in such pipelines. No information on real-world exploitation is available in the provided details.
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2026-15007
Vulnerability details
NVIDIA Megatron-LM contains a vulnerability in inferencing where an Attacker may cause an RCE by convincing a user to load a maliciously crafted input. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, information disclosure,…
more
and data tampering.
- CWE(s)
Related Threats
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
Deserialization vuln (CWE-502) in Megatron-LM inferencing enables RCE when a user loads a malicious crafted input file; directly maps to client-side exploitation (T1203) and user execution of malicious file (T1204.002).
CVEs Like This One
Affected Assets
Mitigating Controls
Mitigating Controls (NIST 800-53 r5) AI
Directly prevents deserialization of untrusted data by requiring validation and sanitization of inputs to NVIDIA Megatron-LM inferencing.
Remediates the specific CWE-502 deserialization flaw in Megatron-LM through timely identification, reporting, and patching.
Enforces least privilege to mitigate privilege escalation and limit damage from RCE triggered by malicious input processing.