# llama.cpp 0.2.0 release adds precise JSON float handling and many platform builds

> The open‑source LLM runtime updates its Jinja float output to match Python and expands binaries for macOS, Linux, Windows, Android and more.

Oossa · 2026-10-14 · https://oossa.com/en/llama-cpp-0-2-0-release-adds-precise-json-float-handling-and-many-platform-build

The ggml‑org team released llama.cpp version b11622 on 2026‑10‑14. The main change fixes how floats are written in Jinja templates: it now uses the shortest decimal that round‑trips, just like Python’s json.dumps. The update also lists a wide range of pre‑built binaries for Apple Silicon, Intel, Linux (CPU, Vulkan, CUDA 12/13, ROCm, OpenVINO, SYCL), Android Snapdragon, Windows (CPU, CUDA, Vulkan, OpenCL) and openEuler.

## The facts

- Release tag b11622 published on Wed Oct 14 2026
- Adds JSON float formatting that matches Python’s json.dumps

## Why it matters

Developers can now rely on accurate numeric values when generating prompts or tool calls, reducing bugs in LLM‑driven applications.

## Sources & references

1. [ggml-org/llama.cpp b11622](https://github.com/ggml-org/llama.cpp/releases/tag/b11622) – llama.cpp, 2026-10-14

Last updated: 2026-10-14