That's the point I am really convinced today that few small size agents actually better than one remote begemoth. I switched fully to local models (from claude, sol, and deepseek). A 27B model can be served on any modern PC or laptop with 64 GB or RAM (or VRAM on non-Macs). A bunch of Qwen coding agents is fantastic with opencode, works great also with pi. The only disadvantage is required compaction due to limited 256k context size, so it stops for few seconds once every hour or so. I am able to produce remarkable things similar to "see what I vibe-coded" posts here on HN but using only open weights, open source, and run everything locally.