GM
A PR got merged today that has been 100% done with local AI (qwen-3.8-flash-next) and nobody complained or even noticed that it was not frontier.
It's a big and active open source project.
I'm p...
I'm running Hermes Agent over qwen3.8:27b on only 32gb of VRAM. It is surprisingly capable but I am still the weak link.
🤔1