CVE-2025-8747 Keras safe_mode bypass allows arbitrary code execution when loading a malicious model.
Attackers can execute arbitrary code by tricking users into loading malicious Keras model files, bypassing safe mode protections.
- Exploited in the wild
- Not known
- Login needed
- Any signed-in user
- Attackable
- Local or same network only
- Victim action needed
- A user has to click or open something
- Fix
- Available
What is CVE-2025-8747?
CVE-2025-8747 is a high-severity unsafe deserialization vulnerability in Google Keras, published on 11 Aug 2025. It can only be exploited from the same machine or local network, and only if a user opens a file or follows a link.
Official description
A safe mode bypass vulnerability in the `Model.load_model` method in Keras versions 3.0.0 through 3.10.0 allows an attacker to achieve arbitrary code execution by convincing a user to load a specially crafted `.keras` model archive.
Weakness: CWE-502
Which versions are affected by CVE-2025-8747?
3 open source packages ship an affected version, including keras. It affects 1 product.
| Package | Affected | Fixed in |
|---|---|---|
keras PyPI | All versions | None |
keras PyPI | From 3.0.0 | 3.11.0 |
keras PyPI | From 3.0.0 | 3.11.0 |
Products include Keras Keras.
How do I fix CVE-2025-8747?
Upgrade keras to 3.11.0 or later.
Upgrade, and only deserialise data from trusted sources or with an allow-list of types. More about Unsafe deserialization.
Is CVE-2025-8747 being exploited?
It is not confirmed as exploited in the wild, but public exploit code for CVE-2025-8747 exists.
Other vulnerabilities in Google Keras
- CVE-2024-3660 Arbitrary code injection vulnerability in Keras framework < 2.13
- CVE-2026-12481 Deserialization of Untrusted Data in keras-team/keras
- CVE-2026-1462 Safe Mode Bypass in keras-team/keras
- CVE-2025-9906 Arbitrary Code execution in Keras Safe Mode
- CVE-2026-11816 Path Traversal in keras-team/keras
Sources
Also known as AZL-66171, GHSA-c9rc-mg46-23w3, PYSEC-2025-75, UBUNTU-CVE-2025-8747.