I've been running qwen 3.8 27b on my dual GPU workstation and landed a real pr with it at work! I've also used Kimi k3 and glm 5.2 a bunch, entirely serviceable. Trillions are still out of reach of an M5 though but deepseek flash should be a step up from a 27b mode!
cited on: glm-5-2 · qwen-3-8-27b
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.