THX-01

by HAL-X AI

A free 322-million-parameter decision model built to run on small machines, with a free API to try it.

Apache-2.0 FreeOpen sourceOpen weightsLocal

View on Hugging Face →

What it does

Choice, yes/no, score, plus extracting numbers, passages and citations from a document
Apache 2.0, weights on Hugging Face, one line to install (pip install thx01)
Results announced by its authors: check them on your own data

! Free alternative to Jev (Apache 2.0), released October 10, 2026, with a free usage-limited API. Its authors say it beats Jev on their own support-ticket benchmark (98.4% vs 97.4%) but note this is not true on every task. These results come from the authors, not an independent test. Information from the authors' Hugging Face article.

Install it

Requirements: Python and pip. A small GPU or a CPU; weights download from Hugging Face.

Install the package

pip install thx01

Load the model

python -c "import thx01; agent = thx01.load('doofz/THX-01')"

The prompt for your agent

Paste it into your coding agent: it checks the requirements and installs it for you.

Claude CodeCodexCursorGemini CLIopencode

Prompt

Install THX-01 (https://huggingface.co/doofz/THX-01) on my machine.
1. Check the requirements: Python and pip. A small GPU or a CPU; weights download from Hugging Face. If anything is missing, tell me before installing anything.
2. Follow the official method:
- Install the package: pip install thx01
- Load the model: python -c "import thx01; agent = thx01.load('doofz/THX-01')"
3. Check that it works, then explain how to launch it and where to start.

Always review what your agent is about to run, especially install commands.