LongJud-Bench - Evidence Marshalling (Second Instance): leaderboard
Metric: Evidence Marshalling Score (%; Claude Sonnet 4.6 judge, 0-10 rubric on the new evidence in the second-instance appeal or response, normalized). Source: arxiv.org. Saturation forecast: Around December 2026. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GLM-4.7 | 71 |
| 2 | Qwen 3.5 Plus | 70 |
| 3 | GPT-5.2 | 63 |
| 4 | DeepSeek V4 Flash | 57 |
| 5 | Qwen 3.5 Flash | 53 |
| 6 | Kimi K2.5 | 52 |
Interactive version: theaggregate.ai/benchmark?slug=longjud-bench-evidence-marshalling-second-instance · How It Works · Data refreshed daily, snapshot 2026-09-25.