LexKairos - Peremptory Period Reasoning: leaderboard
Metric: Accuracy (%; 180 items: decide whether a peremptory period has lapsed under the applicable statute; vanilla setting, the model answers directly without extra reasoning steps (thinking disabled where the model allows it); temperature 0, 8,000 output tokens). Source: arxiv.org. Saturation forecast: Around December 2026. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash (Minimal) | 77.22 |
| 2 | DeepSeek V4 Flash (Non-reasoning) | 62.78 |
| 3 | Qwen 3.5 9B (Non-reasoning) | 62.22 |
| 4 | DeepSeek R1 | 61.67 |
| 5 | GPT-5.4 (Non-reasoning) | 59.44 |
| 6 | Llama 3.1 8B | 40 |
Interactive version: theaggregate.ai/benchmark?slug=lexkairos-peremptory-period-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-29.