CVE-2026-70640 (GCVE-0-2026-70640)
Vulnerability from cvelistv5 – Published: 2026-08-06 15:38 – Updated: 2026-08-14 16:51 X_Open Source
VLAI
EPSS
VEX
Title
llama.cpp b1886–b7445 Race Condition Use-After-Free via llama-android.cpp
Summary
llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Android JNI wrapper where bench_1model() and free_1context() lack synchronization, allowing Thread A to operate on freed memory while Thread B concurrently frees the llama_context. Attackers can exploit this by performing heap spray with attacker-controlled data containing a fake vtable to hijack the vtable pointer at offset +0x30, causing llama_batch_allocr::clear() to dereference arbitrary memory and achieve remote code execution.
Severity
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-08-07 16:18 UTC
CWE
Assigner
References
4 references
| URL | Tags |
|---|---|
| https://github.com/ggml-org/llama.cpp/releases/ta… | release-notes |
| https://github.com/ggml-org/llama.cpp/commit/5c0d… | patch |
| https://github.com/Vladimir-tokarev-cyera/llama-c… | related |
| https://www.vulncheck.com/advisories/llama-cpp-b1… | third-party-advisory |
Impacted products
Date Public
2026-05-06 00:00
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-70640",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-08-07T16:18:29.875187Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-08-07T16:18:40.029Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unaffected",
"product": "llama.cpp",
"vendor": "ggml-org",
"versions": [
{
"lessThanOrEqual": "b7445",
"status": "affected",
"version": "b1886",
"versionType": "custom"
},
{
"lessThanOrEqual": "0.17.1",
"status": "affected",
"version": "0.9.0",
"versionType": "semver"
},
{
"status": "unaffected",
"version": "5c0d18881e0e9794c96b2602736b758bac9d9388",
"versionType": "git"
}
]
}
],
"cpeApplicability": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:*",
"versionEndIncluding": "b7445",
"versionStartIncluding": "b1886",
"vulnerable": true
},
{
"criteria": "cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:*",
"versionEndIncluding": "0.17.1",
"versionStartIncluding": "0.9.0",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "finder",
"value": "Vladimir Tokarev (@G1ND1L4) - Vulnerability Research Tech Lead, Cyera"
},
{
"lang": "en",
"type": "finder",
"value": "Ofek Itach (@ofekitach) - Security Research Team Lead, Cyera"
}
],
"datePublic": "2026-05-06T00:00:00.000Z",
"descriptions": [
{
"lang": "en",
"value": "llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Android JNI wrapper where bench_1model() and free_1context() lack synchronization, allowing Thread A to operate on freed memory while Thread B concurrently frees the llama_context. Attackers can exploit this by performing heap spray with attacker-controlled data containing a fake vtable to hijack the vtable pointer at offset +0x30, causing llama_batch_allocr::clear() to dereference arbitrary memory and achieve remote code execution."
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "HIGH",
"attackRequirements": "NONE",
"attackVector": "LOCAL",
"baseScore": 7.3,
"baseSeverity": "HIGH",
"exploitMaturity": "NOT_DEFINED",
"privilegesRequired": "NONE",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "PASSIVE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:L/AC:H/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "HIGH",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
},
{
"cvssV3_1": {
"attackComplexity": "HIGH",
"attackVector": "LOCAL",
"availabilityImpact": "HIGH",
"baseScore": 7,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.1/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.1"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-476",
"description": "NULL Pointer Dereference",
"lang": "en",
"type": "CWE"
},
{
"cweId": "CWE-362",
"description": "Concurrent Execution using Shared Resource with Improper Synchronization (\u0027Race Condition\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-08-14T16:51:39.415Z",
"orgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
"shortName": "VulnCheck"
},
"references": [
{
"tags": [
"release-notes"
],
"url": "https://github.com/ggml-org/llama.cpp/releases/tag/b7446"
},
{
"tags": [
"patch"
],
"url": "https://github.com/ggml-org/llama.cpp/commit/5c0d18881e0e9794c96b2602736b758bac9d9388"
},
{
"tags": [
"related"
],
"url": "https://github.com/Vladimir-tokarev-cyera/llama-cpp-security-patches"
},
{
"tags": [
"third-party-advisory"
],
"url": "https://www.vulncheck.com/advisories/llama-cpp-b1886-b7445-race-condition-use-after-free-via-llama-android-cpp"
}
],
"source": {
"discovery": "EXTERNAL"
},
"tags": [
"x_open-source"
],
"title": "llama.cpp b1886\u2013b7445 Race Condition Use-After-Free via llama-android.cpp",
"x_generator": {
"engine": "vulncheck"
}
}
},
"cveMetadata": {
"assignerOrgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
"assignerShortName": "VulnCheck",
"cveId": "CVE-2026-70640",
"datePublished": "2026-08-06T15:38:22.994Z",
"dateReserved": "2026-08-04T20:17:18.298Z",
"dateUpdated": "2026-08-14T16:51:39.415Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2",
"vulnerability-lookup:meta": {
"epss": {
"cve": "CVE-2026-70640",
"date": "2026-09-20",
"epss": "0.0016",
"percentile": "0.05563"
},
"nvd": {
"cve": {
"affected": [
{
"affectedData": [
{
"defaultStatus": "unaffected",
"product": "llama.cpp",
"vendor": "ggml-org",
"versions": [
{
"lessThanOrEqual": "b7445",
"status": "affected",
"version": "b1886",
"versionType": "custom"
},
{
"lessThanOrEqual": "0.17.1",
"status": "affected",
"version": "0.9.0",
"versionType": "semver"
},
{
"status": "unaffected",
"version": "5c0d18881e0e9794c96b2602736b758bac9d9388",
"versionType": "git"
}
]
}
],
"source": "disclosure@vulncheck.com"
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Android JNI wrapper where bench_1model() and free_1context() lack synchronization, allowing Thread A to operate on freed memory while Thread B concurrently frees the llama_context. Attackers can exploit this by performing heap spray with attacker-controlled data containing a fake vtable to hijack the vtable pointer at offset +0x30, causing llama_batch_allocr::clear() to dereference arbitrary memory and achieve remote code execution."
}
],
"id": "CVE-2026-70640",
"lastModified": "2026-08-31T20:31:19.777",
"metrics": {
"cvssMetricV31": [
{
"cvssData": {
"attackComplexity": "HIGH",
"attackVector": "LOCAL",
"availabilityImpact": "HIGH",
"baseScore": 7.0,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.1/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.1"
},
"exploitabilityScore": 1.0,
"impactScore": 5.9,
"source": "disclosure@vulncheck.com",
"type": "Secondary"
}
],
"cvssMetricV40": [
{
"cvssData": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "HIGH",
"attackRequirements": "NONE",
"attackVector": "LOCAL",
"availabilityRequirement": "NOT_DEFINED",
"baseScore": 7.3,
"baseSeverity": "HIGH",
"confidentialityRequirement": "NOT_DEFINED",
"exploitMaturity": "NOT_DEFINED",
"integrityRequirement": "NOT_DEFINED",
"modifiedAttackComplexity": "NOT_DEFINED",
"modifiedAttackRequirements": "NOT_DEFINED",
"modifiedAttackVector": "NOT_DEFINED",
"modifiedPrivilegesRequired": "NOT_DEFINED",
"modifiedSubAvailabilityImpact": "NOT_DEFINED",
"modifiedSubConfidentialityImpact": "NOT_DEFINED",
"modifiedSubIntegrityImpact": "NOT_DEFINED",
"modifiedUserInteraction": "NOT_DEFINED",
"modifiedVulnAvailabilityImpact": "NOT_DEFINED",
"modifiedVulnConfidentialityImpact": "NOT_DEFINED",
"modifiedVulnIntegrityImpact": "NOT_DEFINED",
"privilegesRequired": "NONE",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "PASSIVE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:L/AC:H/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
"version": "4.0",
"vulnAvailabilityImpact": "HIGH",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"source": "disclosure@vulncheck.com",
"type": "Secondary"
}
],
"ssvcV203": [
{
"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"ssvcData": {
"id": "CVE-2026-70640",
"options": [
{
"exploitation": "none"
},
{
"automatable": "no"
},
{
"technicalImpact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-08-07T16:18:29.875187Z",
"version": "2.0.3"
}
}
]
},
"published": "2026-08-06T22:18:28.927",
"references": [
{
"source": "disclosure@vulncheck.com",
"url": "https://github.com/Vladimir-tokarev-cyera/llama-cpp-security-patches"
},
{
"source": "disclosure@vulncheck.com",
"url": "https://github.com/ggml-org/llama.cpp/commit/5c0d18881e0e9794c96b2602736b758bac9d9388"
},
{
"source": "disclosure@vulncheck.com",
"url": "https://github.com/ggml-org/llama.cpp/releases/tag/b7446"
},
{
"source": "disclosure@vulncheck.com",
"url": "https://www.vulncheck.com/advisories/llama-cpp-b1886-b7445-race-condition-use-after-free-via-llama-android-cpp"
}
],
"sourceIdentifier": "disclosure@vulncheck.com",
"vulnStatus": "Awaiting Analysis",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-362"
},
{
"lang": "en",
"value": "CWE-476"
}
],
"source": "disclosure@vulncheck.com",
"type": "Secondary"
}
]
}
},
"suse_vex": {
"aggregate_severity": "important",
"current_release_date": "2026-08-07T16:58:44Z",
"cve": "CVE-2026-70640",
"id": "CVE-2026-70640",
"initial_release_date": "2026-08-07T16:58:44Z",
"source": "SUSE CSAF VEX",
"status": "interim",
"title": "SUSE CVE CVE-2026-70640",
"url": "https://ftp.suse.com/pub/projects/security/csaf-vex/cve-2026-70640.json",
"version": "2"
},
"vulnrichment": {
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-70640",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-08-07T16:18:29.875187Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-08-07T16:18:35.311Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unaffected",
"product": "llama.cpp",
"vendor": "ggml-org",
"versions": [
{
"lessThanOrEqual": "b7445",
"status": "affected",
"version": "b1886",
"versionType": "custom"
},
{
"lessThanOrEqual": "0.17.1",
"status": "affected",
"version": "0.9.0",
"versionType": "semver"
},
{
"status": "unaffected",
"version": "5c0d18881e0e9794c96b2602736b758bac9d9388",
"versionType": "git"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "finder",
"value": "Vladimir Tokarev (@G1ND1L4) - Vulnerability Research Tech Lead, Cyera"
},
{
"lang": "en",
"type": "finder",
"value": "Ofek Itach (@ofekitach) - Security Research Team Lead, Cyera"
}
],
"datePublic": "2026-05-06T00:00:00.000Z",
"descriptions": [
{
"lang": "en",
"value": "llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Android JNI wrapper where bench_1model() and free_1context() lack synchronization, allowing Thread A to operate on freed memory while Thread B concurrently frees the llama_context. Attackers can exploit this by performing heap spray with attacker-controlled data containing a fake vtable to hijack the vtable pointer at offset +0x30, causing llama_batch_allocr::clear() to dereference arbitrary memory and achieve remote code execution."
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "HIGH",
"attackRequirements": "NONE",
"attackVector": "LOCAL",
"baseScore": 7.3,
"baseSeverity": "HIGH",
"exploitMaturity": "NOT_DEFINED",
"privilegesRequired": "NONE",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "PASSIVE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:L/AC:H/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "HIGH",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
},
{
"cvssV3_1": {
"attackComplexity": "HIGH",
"attackVector": "LOCAL",
"availabilityImpact": "HIGH",
"baseScore": 7,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.1/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.1"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-476",
"description": "NULL Pointer Dereference",
"lang": "en",
"type": "CWE"
},
{
"cweId": "CWE-362",
"description": "Concurrent Execution using Shared Resource with Improper Synchronization (\u0027Race Condition\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-08-07T19:03:37.562Z",
"orgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
"shortName": "VulnCheck"
},
"references": [
{
"tags": [
"release-notes"
],
"url": "https://github.com/ggml-org/llama.cpp/releases/tag/b7446"
},
{
"tags": [
"patch"
],
"url": "https://github.com/ggml-org/llama.cpp/commit/5c0d18881e0e9794c96b2602736b758bac9d9388"
},
{
"tags": [
"related"
],
"url": "https://github.com/Vladimir-tokarev-cyera/llama-cpp-security-patches"
},
{
"tags": [
"third-party-advisory"
],
"url": "https://www.vulncheck.com/advisories/llama-cpp-b1886-b7445-race-condition-use-after-free-via-llama-android-cpp"
}
],
"source": {
"discovery": "EXTERNAL"
},
"tags": [
"x_open-source"
],
"title": "llama.cpp b1886\u2013b7445 Race Condition Use-After-Free via llama-android.cpp",
"x_generator": {
"engine": "vulncheck"
}
}
},
"cveMetadata": {
"assignerOrgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
"assignerShortName": "VulnCheck",
"cveId": "CVE-2026-70640",
"datePublished": "2026-08-06T15:38:22.994Z",
"dateReserved": "2026-08-04T20:17:18.298Z",
"dateUpdated": "2026-08-07T19:03:37.562Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
}
}
Loading…
Loading…
Experimental. This forecast is provided for visualization only and may change without notice. Do not use it for operational decisions.
Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
Loading…
The MITRE ATT&CK techniques below are AI-generated suggestions, inferred from the description of the
vulnerability by the CIRCL/vulnerability-attack-technique-classification-roberta-base
model, served locally by ML-Gateway.
They have not been verified by an analyst and are provided for guidance only.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Loading…
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.
Loading…