# llama.cpp b12425 adds Qwen3-ForcedAligner support

> The release reads Qwen3-VL’s output-head size from its tensor, allowing llama.cpp to handle Qwen3-ForcedAligner’s smaller timestamp output.

Oossa · 2026-11-15 · https://oossa.com/en/llama-cpp-b12425-adds-qwen3-forcedaligner-support

The llama.cpp project has released version b12425 with a change that allows it to read Qwen3-VL’s output-head size directly from the model tensor. The change supports Qwen3-ForcedAligner, whose output has 5,000 timestamp classes—fewer than the model’s vocabulary.

## The facts

- The change is listed as pull request #32349.
- Version b12425 includes downloadable builds for macOS, Linux and iOS.

## Why it matters

People running Qwen3-ForcedAligner with llama.cpp can use a model whose timestamp output is smaller than its vocabulary; the release notes do not specify setup requirements.

## Sources & references

1. [ggml-org/llama.cpp b12425](https://github.com/ggml-org/llama.cpp/releases/tag/b12425) – llama.cpp, 2026-11-15

Last updated: 2026-11-15
