---
id: CVE-2026-43632
title: >-
  llama.cpp builds b7492 through the latest b9060 contains a use-after-free
  vulnerability in llama-server affecting six tokenization endpoints (/tokenize,
  /detokenize, /infill, /apply-template, /rerank, and /anthropic/count_tokens)
  that by…
summary: >-
  llama.cpp builds b7492 through the latest b9060 contains a use-after-free
  vulnerability in llama-server affecting six tokenization endpoints (/tokenize,
  /detokenize, /infill, /apply-template, /rerank, and /anthropic/count_tokens)
  that by…
severity: high
cvss: 8.1
cvssVector: 'CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H'
cwe:
  - CWE-367
  - CWE-416
vendor: ggml
product: llama.cpp
affected:
  - 'llama.cpp >= b7492, <= b9060'
published: '2026-08-06'
updated: '2026-09-21'
sourceUpdated: '2026-09-21T18:11:15.430'
source: NVD
sourceUrl: 'https://nvd.nist.gov/vuln/detail/CVE-2026-43632'
references:
  - url: 'https://github.com/Vladimir-tokarev-cyera/llama-cpp-security-patches'
    label: disclosure@vulncheck.com
  - url: >-
      https://www.vulncheck.com/advisories/llama-cpp-b7492-b9060-use-after-free-in-tokenization-endpoints
    label: disclosure@vulncheck.com
tags:
  - nvd
epss: 0.00476
epssPercentile: 0.38475
ingestedAt: '2026-08-09T00:31:24.708Z'
---

## Overview

llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in llama-server affecting six tokenization endpoints (/tokenize, /detokenize, /infill, /apply-template, /rerank, and /anthropic/count_tokens) that bypass the task queue and access ctx_server.vocab directly on HTTP worker threads. Attackers can exploit a time-of-check-time-of-use race condition where the main thread destroys and frees vocab after the synchronization lock is released but before the handler finishes using it, causing a crash or potential code execution when --sleep-idle-seconds is configured.

## Affected

- `llama.cpp >= b7492, <= b9060`

## Remediation

Refer to the linked advisories for vendor-supplied fixes and affected version ranges.