CVE-2025-3136VulnerabilityPublished 29 May 2025Updated 8 Sep 2026

CVE-2025-3136: PyTorch CUDACachingAllocator.cpp torch.cuda.memory.caching_allocator_delete memory corruption

Affects pytorchSource GCVE Database

Not known to be exploited : upgrade to 2.7.0 when convenient.

Data as of 25 Oct 2026, 19:02 UTC

Severity

Medium

Rated by Threat Radar · No CVSS score from NVD or the CNA

EPSS: 0.2% chance of exploitation in 30 days – higher than 17% of CVEsExploit Prediction Scoring System: the likelihood attackers will use this flaw soon.
Patch: Available – fixed in 2.7.0

Exploitation evidence

Disclosed: evidence found.
CVE published 29 May 2025
Detectable: no evidence on record.
No public nuclei template indexed
Weaponised: no evidence on record.
No public exploit code indexed in Metasploit or Exploit-DB
Exploited in the wildExploited: no evidence on record.
Not in CISA KEV and no source reports exploitation
Observed by OffSeq MirageSeen by Mirage: no evidence on record.
No attack traffic recorded by OffSeq Mirage sensors

A missing step is not evidence of safety. Sources: cve.org (opens in a new tab), the Metasploit, Exploit-DB and nuclei-templates index as of 2026-10-25.

Disclosed: The CVE record is public on cve.org. Its publication is t = 0.
Detectable: A public nuclei template can find exposed instances, for defenders and attackers alike.
Weaponised: Public exploit code exists in Metasploit or Exploit-DB, or CISA SSVC reports a public proof of concept.
Exploited in the wild: CISA KEV lists it, CISA SSVC rates exploitation active, or a source reports attacks.
Observed by OffSeq Mirage: OffSeq Mirage honeypot sensors recorded attack traffic for this CVE, or found exposed hosts without attack traffic.

Moments are spaced evenly, not to scale; a dashed line marks a gap of six months or more. Deltas count from t = 0. Sources that give only a date (CISA KEV, Exploit-DB, or a time of exactly 00:00Z) read “same day” rather than hours. Metasploit modules carry the flaw’s disclosure date, not their release date, and nuclei templates carry no date, so both appear above but not on the timeline.

What to do

Upgrade pytorch to 2.7.0.

Affected software

Product	Affected	Fixed in
Bitnamipytorch`pkg:bitnami/pytorch`Other threats	Affected >=2.6.0 <2.7.0	Fixed in 2.7.0

Check your own hosts for these packages with threat-finder (opens in a new tab), our free open-source scanner.

AI-suggested

Apply the official patch by upgrading PyTorch to version 2.7.0 or later, as the vulnerability affects versions >=2.6.0 and <2.7.0. Since a patch is available, updating to a fixed version is the recommended remediation.

Description

A memory corruption vulnerability exists in PyTorch version 2.6.0 affecting the function torch.cuda.memory.caching_allocator_delete in the file c10/cuda/CUDACachingAllocator.cpp. The issue requires local access to exploit and has been publicly disclosed. A patch is available to address this vulnerability.

Impact

The vulnerability can lead to memory corruption when the affected function is invoked, which may cause application instability or crashes. Exploitation requires local access, limiting remote attack vectors. No active exploitation has been reported.

Summary

CVE-2025-3136 is a memory corruption vulnerability in PyTorch 2.6.0 specifically within the CUDA memory caching allocator's deletion function. The flaw allows local attackers to manipulate memory, potentially causing corruption. The vulnerability has been publicly disclosed, but there are no known exploits in the wild. A fix has been made available for versions starting from 2.7.0 onward.

References

Community reviews

Loading reviews…

Exploitation evidence

What to do

Affected software

Description

AI analysisMachine-generated, updated 8 Sep 2026, 13:23 UTC

Technical details

References

Community reviews

Related threats