PyTorch 2.6.0 CUDACachingAllocator.cpp torch.cuda.memory.caching_allocator_delete memory corruption
| CVSS Meta Temp Score | Current Exploit Price (≈) | CTI Interest Score |
|---|---|---|
| 3.0 | $0-$5k | 0.00 |
Summary
A vulnerability marked as critical has been reported in PyTorch 2.6.0. This affects the function torch.cuda.memory.caching_allocator_delete of the file c10/cuda/CUDACachingAllocator.cpp. The manipulation leads to memory corruption.
This vulnerability is documented as CVE-2025-3136. The attack needs to be performed locally. Additionally, an exploit exists.
Details
A vulnerability, which was classified as problematic, has been found in PyTorch 2.6.0. This issue affects the function torch.cuda.memory.caching_allocator_delete of the file c10/cuda/CUDACachingAllocator.cpp. The manipulation with an unknown input leads to a memory corruption vulnerability. Using CWE to declare the problem leads to CWE-119. The product performs operations on a memory buffer, but it can read from or write to a memory location that is outside of the intended boundary of the buffer. Impacted is availability.
The advisory is shared at github.com. The identification of this vulnerability is CVE-2025-3136. The exploitation is known to be easy. An attack has to be approached locally. Technical details as well as a public exploit are known.
The exploit is available at github.com. It is declared as proof-of-concept.
There is no information about possible countermeasures known. It may be suggested to replace the affected object with an alternative product.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Product
Name
Version
Website
- Product: https://github.com/pytorch/pytorch/
CPE 2.3
CPE 2.2
CVSSv4
VulDB Vector: 🔍VulDB Reliability: 🔍
CVSSv3
VulDB Meta Base Score: 3.3VulDB Meta Temp Score: 3.0
VulDB Base Score: 3.3
VulDB Temp Score: 3.0
VulDB Vector: 🔍
VulDB Reliability: 🔍
CVSSv2
| AV | AC | Au | C | I | A |
|---|---|---|---|---|---|
| 💳 | 💳 | 💳 | 💳 | 💳 | 💳 |
| 💳 | 💳 | 💳 | 💳 | 💳 | 💳 |
| 💳 | 💳 | 💳 | 💳 | 💳 | 💳 |
| Vector | Complexity | Authentication | Confidentiality | Integrity | Availability |
|---|---|---|---|---|---|
| Unlock | Unlock | Unlock | Unlock | Unlock | Unlock |
| Unlock | Unlock | Unlock | Unlock | Unlock | Unlock |
| Unlock | Unlock | Unlock | Unlock | Unlock | Unlock |
VulDB Base Score: 🔍
VulDB Temp Score: 🔍
VulDB Reliability: 🔍
Exploiting
Class: Memory corruptionCWE: CWE-119
CAPEC: 🔍
ATT&CK: 🔍
Physical: Partially
Local: Yes
Remote: No
Availability: 🔍
Access: Public
Status: Proof-of-Concept
Download: 🔍
EPSS Score: 🔍
EPSS Percentile: 🔍
Price Prediction: 🔍
Current Price Estimation: 🔍
| 0-Day | Unlock | Unlock | Unlock | Unlock |
|---|---|---|---|---|
| Today | Unlock | Unlock | Unlock | Unlock |
Threat Intelligence
Interest: 🔍Active Actors: 🔍
Active APT Groups: 🔍
Countermeasures
Recommended: no mitigation knownStatus: 🔍
0-Day Time: 🔍
Timeline
04/02/2025 🔍04/02/2025 🔍
04/02/2025 🔍
Sources
Product: github.comAdvisory: 149821
Status: Not defined
Confirmation: 🔍
CVE: CVE-2025-3136 (🔍)
GCVE (CVE): GCVE-0-2025-3136
GCVE (VulDB): GCVE-100-303041
scip Labs: https://www.scip.ch/en/?labs.20161013
Entry
Created: 04/02/2025 22:52Changes: 04/02/2025 22:52 (56)
Complete: 🔍
Submitter: Default436352
Cache ID: 216:240:103
Submit
Accepted
- Submit #525252: pytorch pytorch (in torch.cuda.memory.caching_allocator_delete) torch 2.6.0 invalid memory access (by Default436352)
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
No comments yet. Languages: en.
Please log in to comment.