---
id: CVE-2025-62608
title: MLX is an array framework for machine learning on Apple silicon
summary: >-
  MLX is an array framework for machine learning on Apple silicon. Prior to
  version 0.29.4, there is a heap buffer overflow in mlx::core::load() when
  parsing malicious NumPy .npy files. Attacker-controlled file causes 13-byte
  out-of-bounds…
severity: critical
cvss: 9.1
cvssVector: 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H'
cwe:
  - CWE-122
vendor: ml-explore
product: mlx
affected:
  - mlx < 0.29.4
patched:
  - mlx 0.29.4
published: '2025-11-21'
updated: '2026-10-08'
sourceUpdated: '2026-10-08T10:10:00.227'
source: NVD
sourceUrl: 'https://nvd.nist.gov/vuln/detail/CVE-2025-62608'
references:
  - url: 'https://github.com/ml-explore/mlx/pull/1'
    label: security-advisories@github.com
  - url: 'https://github.com/ml-explore/mlx/pull/2'
    label: security-advisories@github.com
  - url: 'https://github.com/ml-explore/mlx/security/advisories/GHSA-w6vg-jg77-2qg6'
    label: security-advisories@github.com
tags:
  - nvd
epss: 0.00527
epssPercentile: 0.42822
ingestedAt: '2026-10-08T10:28:20.513Z'
---

## Overview

MLX is an array framework for machine learning on Apple silicon. Prior to version 0.29.4, there is a heap buffer overflow in mlx::core::load() when parsing malicious NumPy .npy files. Attacker-controlled file causes 13-byte out-of-bounds read, leading to crash or information disclosure. This issue has been patched in version 0.29.4.

## Affected

- `mlx < 0.29.4`

## Remediation

Upgrade past the affected range:

- `mlx 0.29.4`