# llama.cpp adds MiniCPM-OCR 1.0 support in new b11826 release

> The open‑source LLM runner now includes MiniCPM-OCR 1.0 and provides new binaries for macOS, iOS and several Linux builds.

Oossa · 2026-10-22 · https://oossa.com/en/llama-cpp-adds-minicpm-ocr-1-0-support-in-new-b11826-release

The ggml‑org team released llama.cpp version b11826 on Oct 22, 2026. The update adds MiniCPM-OCR 1.0 support, letting the engine run OCR models. Pre‑built binaries are now available for Apple Silicon, Intel macOS, iOS, and Ubuntu on x64, arm64, s390x, plus Vulkan and CUDA‑12 GPU versions. The release links are posted on the GitHub tag page.

## The facts

- Release version b11826 published on 2026-10-22
- Adds MiniCPM-OCR 1.0 support

## Why it matters

Developers can now run OCR tasks locally with llama.cpp without needing separate software.

## Sources & references

1. [ggml-org/llama.cpp b11826](https://github.com/ggml-org/llama.cpp/releases/tag/b11826) – llama.cpp, 2026-10-22

Last updated: 2026-10-22