Interpretive Canons (BVerfG) - Argument Gate: leaderboard
Metric: Positive-class F1 (%; argument gate (argument): whether a candidate sentence supports a given reading at all, sentence-level binary classification in German Federal Constitutional Court decisions; expert-written prompt with worked examples (the same for every model), reasoning models at temperature 1.0; decision-disjoint held-out test split (60% of each subtask, 1:3 positives to negatives)). Source: arxiv.org. Saturation forecast: Around December 2026. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | MiniMax-M3 | 74.7 |
| 2 | DeepSeek V4 Flash | 74 |
| 3 | Gemini 3.5 Flash Lite | 73.2 |
| 4 | DeepSeek V4 Pro | 65.7 |
Interactive version: theaggregate.ai/benchmark?slug=interpretive-canons-bverfg-argument-gate · How It Works · Data refreshed daily, snapshot 2026-09-26.