LongJud-Bench - Evidence Marshalling (First Instance): leaderboard
Metric: Evidence Marshalling Score (%; Claude Sonnet 4.6 judge, 0-10 rubric on the evidence list in the first-instance complaint or defense, normalized). Source: arxiv.org. Saturation forecast: Around December 2026. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Kimi K2.5 | 74 |
| 2 | GLM-4.7 | 72 |
| 3 | GPT-5.2 | 71 |
| 4 | DeepSeek V4 Flash | 69 |
| 5 | Qwen 3.5 Plus | 69 |
| 6 | Qwen 3.5 Flash | 50 |
Interactive version: theaggregate.ai/benchmark?slug=longjud-bench-evidence-marshalling-first-instance · How It Works · Data refreshed daily, snapshot 2026-09-25.