I would like to experiment but before investing loads of money, I do have a MacBook Pro with 32GB RAM, M2 Pro.
Which model would maximize versatility given this hardware?
DeepSeek, Gemma, Qwen? Which model size and quantization?
Focus os mostly on a personal agent (OpenClaw, ZeroClaw etc), followed by a lightweight Claude/ChatGPT replacement. Software development not too important (I may just ask for help writing simple scripts here and there etc)
[link] [comments]