LongJud-Bench - Evidentiary Advocacy (First Instance): leaderboard
Metric: Evidentiary Advocacy Score (%; Claude Sonnet 4.6 judge, 0-10 per trial phase on fact-and-evidence use in the first-instance trial statements, normalized). Source: arxiv.org. Saturation forecast: Around December 2026. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.2 | 61 |
| 2 | Kimi K2.5 | 55 |
| 3 | GLM-4.7 | 53 |
| 4 | Qwen 3.5 Plus | 53 |
| 5 | DeepSeek V4 Flash | 52 |
| 6 | Qwen 3.5 Flash | 46 |
Interactive version: theaggregate.ai/benchmark?slug=longjud-bench-evidentiary-advocacy-first-instance · How It Works · Data refreshed daily, snapshot 2026-09-25.