CVE-2024-58400: llama.cpp before b3701 contains an out-of-bounds read vulnerability in llama_decode_internal and llama_encode_
CVE-2024-58400 is a high-severity flaw (CVSS 8.2) affecting affected software. Below: why it ranks where it does, which versions are affected, what to do, and the exact commands.
llama.cpp before b3701 contains an out-of-bounds read vulnerability in llama_decode_internal and llama_encode_internal that allows remote unauthenticated attackers to read memory by supplying unvalidated token IDs. Attackers can send integer-array prompts to the llama-server /completion endpoint with IDs outside 0 to n_vocab to read embedding memory out of bounds or crash the process.
Why this priority
- High severity
- Remotely exploitable, no login, no user action
No fix is published yet: follow the downgrade or mitigation steps.
Risk facts
| CVSS score | 8.2 (High) |
| Attack vector | Network (remote) |
| Attack complexity | Low |
| Privileges needed | None |
| User interaction | None |
| EPSS exploit chance | n/a |
| CISA exploited list | No |
| Weakness | CWE-129 |
| Published | 2026-10-27 |
| Last modified | 2026-10-27 |
Last updated 2026-10-27 · Sources: CVE Program, NVD, CISA KEV, EPSS and vendor advisories · How we score
Timeline
- Published CVE record published
- Updated Record last modified
Advisories and coverage
Research and news
- Vulners Bulletins: CVE-2024-58400 2026-10-27
Scores and fixes here are computed automatically from those and other public sources. Confirm with the vendor advisory before changing production systems.
Quick answers
How do I fix CVE-2024-58400?
No fixed release is listed yet. Watch the vendor advisory and apply any mitigation shown on this page.
Is CVE-2024-58400 being exploited?
CVE-2024-58400 is not on the CISA Known Exploited Vulnerabilities list at the time of the last update. Check again, as this can change.
What does CVE-2024-58400 affect?
CVE-2024-58400 affects affected software. Severity is high (CVSS 8.2). llama.cpp before b3701 contains an out-of-bounds read vulnerability in llama_decode_internal and llama_encode_internal that allows remote unauthenticated attackers to read memory by supplying unvalidated token IDs. Attackers can send intege
How urgent is CVE-2024-58400?
Priority 34/100. High severity; Remotely exploitable, no login, no user action
Software pages
OpenAI and LLM tools vulnerabilities