Gemma 4 31B

“the dense Gemma4 model seems by far more able and intelligent”

“Gemma 31B (via API) did very well relative to cost”

“i got so used to the speed that i basically stopped using other models. now that they no longer offer it, i find myself reaching for a model a lot less”

“getting real instantaneous answers is absolutely crazy. A very different feeling and UX.”

“Gemma4 31B is not coding model but it is excellent at intent understanding and task execution used in agentic software”

“Personality wise, Gemma4 31B is one of the best options out there if you still need your model to make tool calls”

“Gemma 31b is so lazy and barely make tool calls even if you prompt it”

“gemma 4 31b and if you have enough ram then qwen next is also a great one(use medium reasoning effort), slightly worse language than 31b but beats it almost everywhere else, very refusal-happy though”

“Gemma 31B are nerfed to the max and ONLY useful for useless crap”

“I've tested their internal Gemma 4 31B model vs my external model, and can say that the internal model works better for analyzing projects”

“with Gemma4 31B and its finetunes/merges the exl3 quants perform very well for me (using 4.5bpw and larger)”