StartLux-Decision v1.1 tops Decision Index 0.3.2 with open models that think only when unsure

StartLux says its open 27B Instant decision model ranked first on the Decision Index 0.3.2 Hugging Face leaderboard with a Full score of 69.12 as of October 16, ahead of vLLM-SR d3 at 66.26.

The StartLux-Decision v1.1 family spans eight models from 0.8B to 27B in Instant and AutoThink variants, with weights under CC BY-NC 4.0, which bars commercial use.

The company's own figures include 86.9% pair accuracy on its DecisionPivot benchmark for the 27B AutoThink model.