huggingface/pybot: PyBot: a chatbot written in pure Python by a swarm of agents. No weights, no training, just
A swarm of over 1,000 agents wrote a 640 k‑line Python chatbot that scores 26 % on SmolTalkBench, matching a 360 M‑parameter model.
Hugging Face released PyBot, a chatbot written entirely in Python without any neural network weights or training.
The code was generated by a swarm of more than 1,000 agents that produced and evaluated over 5,000 versions, ending up with about 640,000 lines of rule‑based code.
On the SmolTalkBench benchmark, where an LLM judge compares replies to reference answers, PyBot achieved a 26 % win rate, comparable to the SmolLM‑360M‑Instruct model.
The repository contains 528 Python files, a memory store, a knowledge base and a response composer, all built from if‑else branches, regex patterns and lookup tables.
PyBot runs on Python 3.8 or newer, requires no external packages, and can be started with python -m pybot, offering interactive or single‑prompt modes.
What it means for you
For users seeking free, locally runnable chat solutions, it offers an alternative to the top free chat models like Kimi K3 or GLM‑5.3, which still rely on trained weights.
On FreeLLMArena
From our rankings on 21 Nov 2026, based on Arena's blind votes:
- Kimi K3 (Max): #1 among free chat models
Sources
- huggingface/pybot: PyBot: a chatbot written in pure Python by a swarm of agents. No weights, no training, just rules., GitHub trending (official), 12 Nov 2026
Written automatically from the sources above, then checked by our code: every number and name must appear in them, and no sentence may copy a source. Spotted a mistake? Report it.
Which AI can you use for free today?
Every free AI model, ranked by blind human votes, with the memory and speed it needs on your own GPU.