LexKairos - Bounded Event Retrieval: leaderboard
Metric: Accuracy (%; 594 items: retrieve the case events inside a time window, single- and multi-label; 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 | DeepSeek R1 | 85.69 |
| 2 | GPT-5.4 (Non-reasoning) | 84.68 |
| 3 | Gemini 3 Flash (Minimal) | 74.24 |
| 4 | DeepSeek V4 Flash (Non-reasoning) | 57.24 |
| 5 | Qwen 3.5 9B (Non-reasoning) | 36.36 |
| 6 | Llama 3.1 8B | 35.35 |
Interactive version: theaggregate.ai/benchmark?slug=lexkairos-bounded-event-retrieval · How It Works · Data refreshed daily, snapshot 2026-09-29.