---
id: CVE-2026-31221
title: >-
  pytorch-lightning: PyTorch-Lightning: Arbitrary code execution via insecure
  deserialization of checkpoint files (CVE-2026-31221)
summary: >-
  A flaw was found in PyTorch-Lightning. This vulnerability, categorized as
  insecure deserialization (CWE-502), exists in the checkpoint loading
  mechanism. A remote attacker can exploit this by providing a maliciously
  crafted checkpoint file…
severity: high
cvss: 8
cvssVector: 'CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:U/C:H/I:H/A:H'
cvssSource: vendor
cwe: CWE-502
vendor: Red Hat
product: Red Hat Enterprise Linux AI (RHEL AI) 3
affected:
  - ai_inference_server
  - enterprise_linux_ai_rhel_ai 3
  - openshift_ai_rhoai
published: '2026-05-12'
updated: '2026-09-24'
sourceUpdated: '2026-09-24T17:24:29+00:00'
source: CSAF
sourceUrl: 'https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-31221.json'
references:
  - url: >-
      https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-31221.json
  - url: 'https://access.redhat.com/security/cve/CVE-2026-31221'
  - url: 'https://bugzilla.redhat.com/show_bug.cgi?id=2476502'
  - url: 'https://www.cve.org/CVERecord?id=CVE-2026-31221'
  - url: 'https://nvd.nist.gov/vuln/detail/CVE-2026-31221'
  - url: 'https://github.com/Lightning-AI/pytorch-lightning'
  - url: 'https://www.notion.so/CVE-2026-31221-35d1e1393188815f8db7c4fd08076639'
  - url: 'https://pypi.org/project/pytorch-lightning'
  - url: 'https://github.com/advisories/GHSA-75m9-98v2-hjpm'
tags:
  - csaf
  - vex
  - red-hat
  - osv
  - pip
epss: 0.00554
epssPercentile: 0.43881
aliases:
  - GHSA-75m9-98v2-hjpm
  - PYSEC-2026-3043
  - PYSEC-2026-3972
ecosystem: pip
scores:
  vendor: 8
  osv: 7.8
ingestedAt: '2026-07-13T18:57:54.195Z'
---

## Overview

A flaw was found in PyTorch-Lightning. This vulnerability, categorized as insecure deserialization (CWE-502), exists in the checkpoint loading mechanism. A remote attacker can exploit this by providing a maliciously crafted checkpoint file, which, when loaded by the LightningModule.load_from_checkpoint() method, allows the deserialization of arbitrary Python objects. This can lead to arbitrary code execution on the victim's system.

## Vendor advisories

- **Red Hat VEX** · Important · affected: Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI) 3, Red Hat OpenShift AI (RHOAI) · no fix planned: Red Hat AI Inference Server, Red Hat OpenShift AI (RHOAI), Red Hat Enterprise Linux AI (RHEL AI) 3 · updated 2026-09-24 · [vex](https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-31221.json)

**pytorch-lightning: PyTorch-Lightning: Arbitrary code execution via insecure deserialization of checkpoint files** — rated Important by Red Hat. Released 2026-05-12, updated 2026-09-24.

Affected:

- Red Hat AI Inference Server
- Red Hat Enterprise Linux AI (RHEL AI) 3
- Red Hat OpenShift AI (RHOAI)

No fix planned:

- Red Hat AI Inference Server
- Red Hat OpenShift AI (RHOAI)
- Red Hat Enterprise Linux AI (RHEL AI) 3

## Remediation

Will not fix

Workarounds / mitigations:

- To mitigate this issue, avoid loading PyTorch-Lightning checkpoint files from untrusted or unverified sources. Ensure that all checkpoint files originate from trusted environments to prevent the deserialization of malicious Python objects and subsequent arbitrary code execution.

## Package advisory (CVE-2026-31221)

Affected packages:

- `pytorch-lightning <= 2.6.0`

Source: https://osv.dev/vulnerability/GHSA-75m9-98v2-hjpm