large model evaluation

LocalLLaMA Community Proposes a “Good Enough” Model Evaluation Standard

A LocalLLaMA proposal sets task success and cost as model criteria, shifting AI competition beyond benchmark scores.