# llama.cpp update adds JSON depth limit to prevent crashes

> The open‑source LLM runtime now rejects JSON nested deeper than 128 levels, avoiding server crashes on macOS and Linux.

Oossa · 2026-10-25 · https://oossa.com/en/llama-cpp-update-adds-json-depth-limit-to-prevent-crashes

The ggml‑org/llama.cpp project released a fix on Oct 25, 2026. It adds a check that stops processing JSON deeper than 128 levels and returns a 400 error. Previously, extremely nested JSON could overflow the HTTP thread stack and crash the llama‑server (SIGBUS on macOS, SIGSEGV on Linux). The change applies to chat inputs, tool calls, and the /v1/systemone endpoint. The fix was co‑authored by Xuan Son Nguyen of Hugging Face.

## The facts

- Release date: 2026-10-25 ("Sun Oct 25 2026")
- JSON depth limit set to 128 levels

## Why it matters

Developers using llama.cpp can now rely on more stable server behavior when handling complex JSON payloads.

## Sources & references

1. [ggml-org/llama.cpp b11871](https://github.com/ggml-org/llama.cpp/releases/tag/b11871) – llama.cpp, 2026-10-25

Last updated: 2026-10-25