What should we work on?
Models
Download open models from Hugging Face and run them locally with Ollama
Qwen3-Coder-30B-A3B-Instruct
18.6 GB Fits
A mixture-of-experts coding model with 3B active parameters. Fast on Apple silicon.
gpt-oss-20b
12.1 GB Fits
OpenAI's open-weight reasoning model, sized for a laptop.
gemma-3-27b-it
16.5 GB Fits
Google's instruction-tuned Gemma 3 with image input.
Llama-3.3-70B-Instruct
42.5 GBWon't fit
Meta's 70B instruction model.