{"id":"CVE-2025-71439","aliases":["DEBIAN-CVE-2025-71439","EUVD-2025-211201","GHSA-4qpr-vwfh-qwp4","GHSA-g4cc-763q-h9h6","UBUNTU-CVE-2025-71439"],"status":"published","title":"llama.cpp before b5649 Heap Out-of-Bounds Read via llama_vocab::impl::print_info","description":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl::print_info() that allows attackers to crash model loading via oversized special token IDs. Attackers can supply a crafted GGUF model with a short tokenizer.ggml.tokens array, causing llama-cli or llama-server to read past id_to_token and crash.","severity":{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"medium","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-12","score":0.00121,"percentile":0.01958,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-125"],"published":"2026-10-27T00:33:57Z","modified":"2026-11-05T15:22:09Z","fixes":[{"ecosystem":"Debian:14","package":"llama.cpp","fixed_in":["5713+dfsg-1"],"sources":["osv"]}],"affected":[{"source":"osv","ecosystem":"Debian:14","package":"llama.cpp","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"5713+dfsg-1"}]}],"versions":["5151+dfsg-1~exp2","5151+dfsg-1~exp3","5318+dfsg-1","5318+dfsg-2"]},{"source":"osv","ecosystem":"Ubuntu:26.10","package":"llama.cpp","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"}]}],"versions":["8681+dfsg-1","9721+dfsg-1ubuntu1","10438+dfsg-1","10438+dfsg-1ubuntu1"]},{"source":"osv","ecosystem":"Ubuntu:26.04:LTS","package":"llama.cpp","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"}]}],"versions":["5882+dfsg-2","8064+dfsg-1ubuntu1","8681+dfsg-1"]}],"products":[{"vendor":"ggml-org","product":"llama.cpp"}],"references":[{"url":"https://github.com/ggml-org/llama.cpp/security/advisories/GHSA-g4cc-763q-h9h6","tags":["vendor-advisory","exploit","advisory"]},{"url":"https://github.com/ggml-org/llama.cpp/pull/14145","tags":["patch","issue-tracking","fix","report"]},{"url":"https://github.com/ggml-org/llama.cpp/commit/c33fe8b8c4427202706b1434e9fc8ab5752c9cac","tags":["patch","fix","report"]},{"url":"https://github.com/ggml-org/llama.cpp/blob/b5648/src/llama-vocab.cpp#L2771","tags":["technical-description","article"]},{"url":"https://github.com/ggml-org/llama.cpp/releases/tag/b5649","tags":["release-notes","advisory","report"]},{"url":"https://www.vulncheck.com/advisories/llama-cpp-before-b5649-heap-out-of-bounds-read-via-llama-vocab-impl-print-info","tags":["third-party-advisory","advisory","report"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2025-71439","tags":["advisory"]},{"url":"https://github.com/advisories/GHSA-4qpr-vwfh-qwp4"},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2025/71xxx/CVE-2025-71439.json","tags":["advisory"]},{"url":"https://security-tracker.debian.org/tracker/CVE-2025-71439","tags":["advisory"]},{"url":"https://ubuntu.com/security/CVE-2025-71439","tags":["report"]},{"url":"https://www.cve.org/CVERecord?id=CVE-2025-71439","tags":["report"]},{"url":"https://github.com/cveproject/cvelistv5/tree/main/cves/2025/71xxx/cve-2025-71439.json","tags":["report"]},{"url":"https://github.com/ggml-org/llama.cpp/blob/b5648/src/llama-vocab.cpp#l2771","tags":["report"]},{"url":"https://github.com/ggml-org/llama.cpp/security/advisories/ghsa-g4cc-763q-h9h6","tags":["report"]},{"url":"https://nvd.nist.gov/vuln/detail/cve-2025-71439","tags":["report"]}],"sources":{"cve":[{"id":"CVE-2025-71439","title":"llama.cpp before b5649 Heap Out-of-Bounds Read via llama_vocab::impl::print_info","description":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl::print_info() that allows attackers to crash model loading via oversized special token IDs. Attackers can supply a crafted GGUF model with a short tokenizer.ggml.tokens array, causing llama-cli or llama-server to read past id_to_token and crash.","status":"published","published":"2026-10-27T00:33:57.956Z","modified":"2026-11-05T15:22:09.001Z","cvss":[{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"medium","assessor":"cna"},{"version":"3.1","score":5.5,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H","level":"medium","assessor":"cna"}],"cwes":["CWE-125"],"updated_at":"2026-11-05T16:36:33.176634Z"}],"euvd":[{"id":"EUVD-2025-211201","description":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl::print_info() that allows attackers to crash model loading via oversized special token IDs. Attackers can supply a crafted GGUF model with a short tokenizer.ggml.tokens array, causing llama-cli or llama-server to read past id_to_token and crash.","status":"published","published":"2026-10-27T00:33:57Z","modified":"2026-11-05T15:22:09Z","cvss":[{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"medium"}],"updated_at":"2026-11-05T16:59:57.067043Z"}],"ghsa":[{"id":"GHSA-4qpr-vwfh-qwp4","title":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl:...","description":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl::print_info() that allows attackers to crash model loading via oversized special token IDs. Attackers can supply a crafted GGUF model with a short tokenizer.ggml.tokens array, causing llama-cli or llama-server to read past id_to_token and crash.","status":"published","published":"2026-10-27T03:32:10Z","modified":"2026-10-27T03:32:11Z","cvss":[{"version":"3.1","score":5.5,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H","level":"medium"},{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/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","level":"medium"}],"cwes":["CWE-125"],"updated_at":"2026-11-02T15:14:28.078787Z"}],"nvd":[{"id":"CVE-2025-71439","description":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl::print_info() that allows attackers to crash model loading via oversized special token IDs. Attackers can supply a crafted GGUF model with a short tokenizer.ggml.tokens array, causing llama-cli or llama-server to read past id_to_token and crash.","status":"published","published":"2026-10-27T01:19:03.47Z","modified":"2026-11-05T16:17:22.12Z","cvss":[{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/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","level":"medium","assessor":"secondary"},{"version":"3.1","score":5.5,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H","level":"medium","assessor":"secondary"}],"cwes":["CWE-125"],"updated_at":"2026-11-05T19:56:07.634341Z"}],"osv":[{"id":"CVE-2025-71439","title":"llama.cpp before b5649 Heap Out-of-Bounds Read via llama_vocab::impl::print_info","description":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl::print_info() that allows attackers to crash model loading via oversized special token IDs. Attackers can supply a crafted GGUF model with a short tokenizer.ggml.tokens array, causing llama-cli or llama-server to read past id_to_token and crash.","status":"published","published":"2026-10-27T00:33:57.956Z","modified":"2026-11-11T13:20:03.632682022Z","cvss":[{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"medium"}],"cwes":["CWE-125"],"updated_at":"2026-11-07T03:07:58.953998Z"},{"id":"DEBIAN-CVE-2025-71439","description":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl::print_info() that allows attackers to crash model loading via oversized special token IDs. Attackers can supply a crafted GGUF model with a short tokenizer.ggml.tokens array, causing llama-cli or llama-server to read past id_to_token and crash.","status":"published","published":"2026-10-27T01:19:03.47Z","modified":"2026-11-06T10:04:02.454330759Z","cvss":[{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/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","level":"medium"}],"updated_at":"2026-11-06T10:38:22.758306Z"},{"id":"UBUNTU-CVE-2025-71439","description":"llama.cpp before b5649 contains a heap out-of-bounds read vulnerability in llama_vocab::impl::print_info() that allows attackers to crash model loading via oversized special token IDs. Attackers can supply a crafted GGUF model with a short tokenizer.ggml.tokens array, causing llama-cli or llama-server to read past id_to_token and crash.","status":"published","published":"2026-10-27T01:19:00Z","modified":"2026-10-28T20:16:32.703324441Z","cvss":[{"version":"3.1","score":5.5,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H","level":"medium"},{"version":"3.1","score":5.5,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H","level":"medium"},{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"medium"},{"version":"4.0","score":6.8,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"medium"}],"updated_at":"2026-11-02T16:55:55.926217Z"}]}}