Decisions As Next-Token Estimates: DANTE v1, a 4B decision model, and its llama.cpp builds
A class-leading 4B decision model with calibrated odds