We just switched our free coding agent to use GLM 5.1 instead of MiniMax M2.5 Eval performance went from 68% to 83% 🔥 P.S. It's still 100% free npm install -g freebuff

@jahooma Freebuff getting huge 💪

@CodebuffAI Went to the (RL) gym for 3 months, came back huge (in parameters)