# llama.cpp 0.7 adds i32 OpenCL support and new binaries

> The open‑source llama.cpp library released version b11740 on Oct 19 2026, adding 32‑bit integer OpenCL operations and fresh builds for macOS, iOS and Linux.

Oossa · 2026-10-19 · https://oossa.com/en/llama-cpp-0-7-adds-i32-opencl-support-and-new-binaries

The ggml‑org team posted llama.cpp version b11740 on 2026‑10‑19. The update adds OpenCL support for 32‑bit integer GET_ROWS and copy conversions, expanding the library’s GPU capabilities. It also ships pre‑built binaries for macOS Apple Silicon, Intel Macs, iOS, and several Ubuntu variants including CPU, Vulkan and CUDA builds.

## The facts

- Release date: 2026‑10‑19 ("Mon Oct 19 2026 21:58:50 GMT+0200")
- New OpenCL features: i32 GET_ROWS and i32/f32 copy conversions

## Why it matters

Developers can now run llama.cpp models on more GPUs and get ready‑to‑use binaries for their platforms.

## Sources & references

1. [ggml-org/llama.cpp b11740](https://github.com/ggml-org/llama.cpp/releases/tag/b11740) – llama.cpp, 2026-10-19

Last updated: 2026-10-19