StartLux Decision v1.1 Release - DecisionPivot Core Pair Accuracy: leaderboard
DecisionPivot core track: 7,800 program-generated items in 13 families; a contrast group counts only when every member is right (card names no DecisionPivot version; its 27B rows equal the decisionpivot_paper v1.1 values); card: not used for any data, checkpoint or configuration choice 7,800 core items, 13 families Metric: pair accuracy (every member of a contrast group right).
Metric: pair accuracy (every member of a contrast group right). Source: huggingface.co. Saturation forecast: Rough model projection: around 2026. 5 models tracked. Updated 14 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | StartLux-v1.1-4B-AutoThink | 59.9 | |
| 2 | StartLux-v1.1-9B-Instant | 47.5 | |
| 3 | StartLux-v1.1-4B-Instant | 41.5 | |
| 4 | StartLux-v1.1-2B-Instant | 21.1 | |
| 5 | StartLux-v1.1-0.8B-Instant | 8.1 |
Interactive version: theaggregate.ai/benchmark?slug=startlux-decision-v1-1-release-decisionpivot-core-pair-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-16.